DEVELOPMENTAL MEDICINE & CHILD NEUROLOGY ORIGINAL ARTICLE

Prevalence of specific gait abnormalities in children with

cerebral palsy revisited: influence of age, prior surgery, and Gross
Motor Function Classification System level
SUSAN A RETHLEFSEN 1 | GIDEON BLUMSTEIN 2 | ROBERT M KAY 1,2 | FREDERICK DOREY3 | TISHYA A L WREN1,2
1 Children’s Orthopedic Center, Children’s Hospital, Los Angeles, CA; 2 Department of Orthopedic Surgery, Keck School of Medicine of USC, Los Angeles, CA;
3 Department of Orthopedic Surgery, UCLA School of Medicine, Los Angeles, CA, USA.
Correspondence to Susan A Rethlefsen at Children’s Orthopedic Center, Children’s Hospital Los Angeles, 4650 Sunset Blvd, M/S 69, Los Angeles, CA 90027, USA.
E-mail: srethlefsen@chla.usc.edu

PUBLICATION DATA AIM To examine the impact of age, surgery, and Gross Motor Function Classification System
Accepted for publication 8th June 2016. (GMFCS) level on the prevalence of gait problems in children with cerebral palsy (CP).
Published online METHOD Gait analysis records were retrospectively reviewed for ambulatory patients with
CP. Gait abnormalities were identified using physical exam and kinematic data. Relationships
among age, sex, previous surgery, GMFCS level, and prevalence of gait abnormalities
associated with crouch and out-toeing, and equinus and in-toeing were assessed using
univariable and multivariable logistic regression.
RESULTS One-thousand and five records were reviewed. The most common gait problems
were in-toeing, excessive knee flexion, stiff knee, hip flexion, internal rotation, adduction, and
equinus (all >50%). Odds ratios (OR) for various gait problems associated with crouch and
out-toeing increased (OR 1.07–1.32), and those associated with equinus and in-toeing
decreased (OR 0.80–0.94) significantly with increasing age for patients in GMFCS levels I to
III. The same trends were seen with prior surgery (OR for crouch and out-toeing: 1.86–7.14;
OR for equinus and in-toeing: 0.16–0.59).
INTERPRETATION The prevalence of gait abnormalities varies by GMFCS level, but similarities
exist among levels. The study results suggest that in younger children, particularly those in
GMFCS levels III and IV, treatments for equinus and in-toeing should be undertaken with
caution because these problems tend to decrease with age even without orthopedic
intervention. Such children may end up with the ‘opposite’ deformities of calcaneal crouch
and out-toeing, which tend to increase in prevalence with age.

The prevalence of specific gait abnormalities in children
with cerebral palsy (CP) varies with distribution of involve-
ment (hemiplegia, diplegia, or quadriplegia) and changes
with age and orthopedic surgery.1 A previous study of 492
patients with CP conducted at the authors’ institution
found that patients with hemiplegia had a higher preva-
lence of varus foot deformities when compared with
patients with bilateral involvement. Diplegic gait was char-
acterized by stiff knee in swing, excessive knee flexion, and
in-toeing, with excessive hip flexion and equinus. Patients
with quadriplegia had similar gait problems, with the addi-
tion of excessive hip adduction and a higher prevalence of
excessive knee flexion, scissoring, and valgus foot deformi-
ties than children with diplegia. Rotational malalignment
between the femur and tibia, calcaneal gait, out-toeing,
varus/valgus foot deformities, and hip internal rotation
increased in prevalence with age, while stiff knee gait, out-
toeing, calcaneus, and excessive knee flexion increased in
prevalence when prior surgery had been done. These
findings suggest that calcaneus, excessive knee flexion, and
out-toeing gait are associated with increasing age,
increased level of disability, and prior surgical intervention.
In recent years, the Gross Motor Function Classification
System (GMFCS) has become the criterion standard for
functional classification in children with CP.2–4 To date,
only one study has examined the relationship of specific
gait abnormalities to GMFCS level. The study by Ounpuu
et al.5 included 292 patients with bilateral CP functioning
in GMFCS levels I to III. The authors found that gait
deviations, such as knee flexion at initial contact, were
prevalent in GMFCS levels I to III, with severity increas-
ing as GMFCS level increased. However, there was consid-
erable variability and overlap in the prevalence and severity
of gait problems within and between GMFCS levels, which
suggests that the GMFCS should not be used as a substi-
tute for individualized gait analysis when planning treat-
ment interventions. Nevertheless, knowledge of the
patterns of gait abnormalities particular to a given child’s

© 2016 Mac Keith Press DOI: 10.1111/dmcn.13205 1

GMFCS level may allow clinicians to have more informed
discussions with patients and families about ambulatory
potential, and may allow them to focus their care to maxi-
mize potential outcome.
The aim of the current study was to investigate the rela-
tionships between age, prior surgery, GMFCS level, and
the prevalence of gait abnormalities in a large group of
patients with CP, including children with both unilateral
and bilateral CP functioning in GMFCS levels I to IV.
Based on prior research we hypothesized that the preva-
lence of gait problems associated with crouch and out-
toeing would increase, while those associated with equinus
and in-toeing would decrease with increasing age, increas-
ing GMFCS level (increasing level of disability), and in
cases where prior surgery had been done.
METHOD
Institutional Review Board approval was obtained prior to
commencement of the study. An internal database main-
tained on an ongoing basis in our gait analysis laboratory
was queried retrospectively to identify patients with CP who
underwent pretreatment gait analysis testing in our labora-
tory between January 1993 and February 2014. All partici-
pants had undergone gait analysis testing using a VICON
three-dimensional motion analysis system (Vicon Motion
Systems, Oxford, UK). This system uses a set of 15 to 19
passive retro-reflective markers attached over specific bony
landmarks of the pelvis and lower extremities.
Participants made three or more passes down a 15m
path at a self-selected speed with the markers in place, with
assistive devices if needed. Kinematic data from 8 to 10
representative steps were averaged,6 and the averaged data
were included in the gait analysis report. After each gait
report was completed, an experienced gait laboratory phys-
ical therapist (one of four) reviewed the gait data and iden-
tified gait abnormalities based on specified definitions
(Table SI, online supporting information).1
Each gait problem was coded ‘1’ if the deviation was
present on one or both sides, and ‘0’ if the deviation was
not present on either side. The therapist then entered the
codes into the database. Though some participants had
more than one gait analysis test done and, therefore, more
than one entry in the database, only data from the first gait
analysis session was included for each participant. Informa-
tion about orthopedic surgeries done prior to gait analysis
testing (including soft tissue, bony surgeries, and selective
dorsal rhizotomy) was obtained by patient/parent report
and/or from patients’ medical records. This information
was also entered into the database.
Gait deviations were grouped according to their associa-
tion with equinus and in-toeing, crouch and out-toeing, or
other deviations. Deviations associated with equinus and
in-toeing included equinovarus, equinus, pes varus, recur-
vatum, hip internal rotation, in-toeing, and internal tibial
torsion. Those associated with crouch and out-toeing
included calcaneus, pes valgus, excessive knee flexion,
excessive hip flexion, out-toeing, external tibial torsion,
2 Developmental Medicine & Child Neurology 2016
What this paper adds
• Prevalence of gait problems varies by Gross Motor Function Classification
System level.
• There is a higher prevalence of deviations with increasing ambulatory
disability.
• Likelihood of gait problems associated with crouch and out-toeing increases
with age.
• Likelihood of the same problems also increases when prior surgery has been
done.
• Likelihood of gait problems associated with equinus and in-toeing decreases
with age.
and rotational malalignment. Deviations not thought to be
associated with any particular gait problem grouping
included stiff knee, hip adduction, anterior pelvic tilt, and
posterior pelvic tilt.
Statistical analysis was performed using Stata/IC 14.0
(StataCorp, College Station, TX, USA). Age at surgery
was compared among GMFCS levels and between patients
with and without surgery using the Kruskal–Wallis and
Mann–Whitney rank sum tests. The likelihood of having
each gait abnormality was first assessed using univariable
logistic regression to examine the effects of age, sex, previ-
ous surgery, and GMFCS level separately. Multivariable
logistic regression, which allows for simultaneous evalua-
tion of all variables of interest because they may affect each
other, was then performed to adjust for covariates within
each GMFCS level. The factors included were selected
based on previous studies in the literature. Model fit was
examined using the Hosmer–Lemeshow Goodness of Fit
statistic and the area under the receiver operating charac-
teristic curve. In most cases age was modeled as continu-
ous, but in some cases age was divided into ordinal groups
to provide better goodness of fit. Odds ratios (ORs) and
significance were similar for both models; therefore, results
with age as a continuous variable are presented for consis-
tency. The significance level was set at p<0.05. Gait abnor-
mality variables were divided into three groups for
presentation: equinus and in-toeing, crouch and out-toe-
ing, and other deviations.
RESULTS
One-thousand and five patients were included: 282 func-
tioning in GMFCS level I; 320 in level II; 289 in level III;
and 114 patients in level IV. The age range of participants
at the time of gait analysis was 3 to 21 years (median 9y;
interquartile range 5y). There were 589 males (59%), and
416 females (41%) (Table I). Thirty-two percent of partic-
ipants (324/1005) had undergone orthopedic surgery or
rhizotomy prior to gait analysis testing (Table II).
The average age of participants did not differ signifi-
cantly among the various GMFCS levels (p=0.843). How-
ever, participants who had undergone prior surgery were
significantly older than those who had not (p<0.001)
(Table III). This was true for all participants combined as
well as for each GMFCS level individually.
The prevalence of various gait abnormalities for patients
in each GMFCS level (I–IV) is shown in Figure 1 (see also
Table SII, online supporting information). The most com- Univariable analysis
mon gait deviations seen in patients functioning in Equinus and in-toeing
GMFCS level I were in-toeing (73%) and equinus (51%). The odds of having deviations associated with equinus and
In GMFCS level II, the most prevalent deviations were in- in-toeing decreased with age (including equinus, internal
toeing (73%), hip internal rotation (65%), excessive knee tibial torsion, and in-toeing), when prior surgery had been
flexion (61%), stiff knee (60%), equinus (53%), and exces- done (equinovarus, internal tibial torsion, in-toeing, and
sive hip flexion (51%). The same deviations were most equinus), and with increasing GMFCS level (including
prevalent in GMFCS level III (stiff knee 83%, excessive recurvatum and pes varus). Age had a similar effect on
knee flexion 81%, excessive hip flexion 79%, in-toeing equinus (OR 86, 95% confidence interval [CI] 0.83–0.89),
68%, hip internal rotation 59%), with the addition of in-toeing (OR 0.85, 95% CI 0.82–0.89), and internal tibial
excessive hip adduction (60%). In GMFCS level IV, the torsion (OR 0.85, 95% CI 0.81–0.89). Prior surgery had
most common deviations were excessive knee flexion the greatest impact on equinus (OR 0.40, 95% CI 0.31–
(85%), excessive hip flexion (82%), stiff knee (75%), in- 0.53) and in-toeing (OR 0.44, 95% CI 0.33–0.58)
toeing (62%), hip adduction (60%), and pes valgus (50%) (Table SIII, online supporting information).
(Fig. 1 [or Fig. S1, online supporting information],
Table SII). Crouch and out-toeing
The odds of having gait problems associated with crouch
and out-toeing increased with age (including rotational
malalignment, external tibial torsion, out-toeing, pes val-
Table I: Participant demographics by GMFCS level gus, calcaneus, excessive hip flexion, and excessive knee
flexion), when prior surgery had been done (out-toeing,
GMFCS Participants, % of partic- Age, med- Sex, n
level n ipants ian (IQR) (%) male rotational malalignment, external tibial torsion, calcaneus,
pes valgus, excessive hip flexion, excessive knee flexion, and
I 282 28 9 (6) 172 (61)
excessive hip adduction), and with increasing GMFCS level
II 320 32 9 (5) 187 (58)
III 289 29 9 (5) 152 (53) (including excessive hip flexion, excessive knee flexion, pes
IV 114 11 9 (5) 78 (68) valgus, calcaneus, out-toeing, rotational malalignment, and
All 1005 100 9 (5) 589 (59)
external tibial torsion). Age had the greatest effect on rota-
participants
tional malalignment (OR 1.23, 95% CI 1.18–1.29) and
GMFCS, Gross Motor Function Classification System; IQR, external tibial torsion (OR 1.19, 95% CI 1.13–1.26), while
interquartile range.
prior surgery had the greatest effect on out-toeing (OR
4.35, 95% CI 3.20–5.90) and rotational malalignment (OR
3.70, 95% CI 2.60–5.30) (Table SIII).
Table II: Prior surgical procedures undergone by 324/1005 participants
(32%) Other deviations
Procedure n
The odds of having stiff knee (OR 3.10, 95% CI 2.27–
4.20) and excessive hip adduction (OR 1.35, 95% CI 1.04–
Selective dorsal rhizotomy 13 1.76) increased significantly with prior surgery. The odds
Femoral/Varus derotational osteotomy 46
Tibial derotational osteotomy 22 of having these problems also increased with GMFCS level
Hip adductor lengthening 162 (Table SIII).
Psoas recession 38
Hamstring lengthening 228
Rectus femoris transfer 35 Multivariable analysis
Triceps surae lengthening 214 Equinus and in-toeing
Foot surgery (bony) 23 Controlling for sex and prior surgery, the odds of having
Foot surgery (soft tissue) 35
Unknown 33 equinus and in-toeing decreased by small but statistically
significant amounts with increasing age in GMFCS levels

Table III: Comparison of age between participants with and without prior surgery, by GMFCS level

No prior surgery Prior surgery

GMFCS level Participants, n % Age, median (IQR) Participants, n % Age, median (IQR) p

I 223 79 8 (4) 59 21 13 (5) <0.001

II 231 72 8 (5) 89 28 12 (5) <0.001
III 152 53 7 (3) 137 47 12 (6) <0.001
IV 75 66 8 (4) 39 34 12 (5) <0.001
All participants 681 68 8 (5) 324 32 12 (6) <0.001

GMFCS, Gross Motor Function Classification System; IQR, interquartile range.

Prevalence of Gait Abnormalities in CP Susan A Rethlefsen et al. 3

 Equinus and in-toeing
80
70

Prevalence (%)
60
50
40 GMFCS I
30
GMFCS II
20
10 GMFCS III
0
GMFCS IV

us

us
m

us
n

on
in

io

tu
in

r
ar
si
e

at

va
va
u

or
to

ov
ot

Eq
All

s
-

ur
lt

in
lr
In

Pe
ia

ec

u
na

ib

Eq
R
r

lt
te

na
in

r
ip

te
H

In
Crouch and out-toeing
90
80
Prevalence (%)

70
60
50
40 GMFCS I
30
20 GMFCS II
10
0 GMFCS III
us

us

ng
n

on
t
on

en
o

lg

ne

ei
xi

si
xi

nm
GMFCS IV
fle

fle

va

or
-to
ca

lt
ig
ut
s
ee

ip

al

ia
al
Pe

O
H

C
kn

tib
al
All
lm
ive

al
rn
na
ss

te
io
ce

Ex
at
Ex

ot
R

Other deviations
90
80
70
Prevalence (%)

60
50 GMFCS I
40
30 GMFCS II
20
GMFCS III
10
0 GMFCS IV
ee

t
til

til
io
kn

c
ct

All
vi

vi
du
iff

el

el
St

ad

rp

rp
rio

io
ip

er
H

te

st
An

Po

Figure 1: Prevalence of gait abnormalities by Gross Motor Function Classification System (GMFCS) level.

I, II, and III (equinus, internal tibial torsion, in-toeing, GMFCS level III (OR 0.50, 95% CI 0.29–0.86), and by
plus internal hip rotation for GMFCS level II). The largest 72% in level IV (OR 0.28, 95% CI 0.10–0.78) with prior
decrease in prevalence due to age was seen for internal tib- surgery. Odds of in-toeing were reduced by 44% at
ial torsion in GMFCS level III (OR 0.80, 95% CI 0.71– GMFCS level III (OR 0.56, 95% CI 0.32–0.99), and by
0.89). Age did not affect the odds of having equinus and 66% at level IV (OR 0.34, 95% CI 0.14–0.82) with prior
in-toeing in GMFCS level IV (Fig. 2; Table SIV, online surgery. Prior surgery did not affect the odds of having
supporting information). equinus and in-toeing in GMFCS levels I and II (Fig. 2;
Controlling for age and sex, the odds of having equinus Table SV, online supporting information).
and in-toeing also decreased significantly in GMFCS levels
III (hip internal rotation, in-toeing, equinus) and IV (in- Crouch and out-toeing
toeing, equinus, equinovarus) when prior surgery had been Controlling for sex and prior surgery, the odds of having
done. Odds of having equinus were lowered by 50% in crouch and out-toeing increased with increasing age in

4 Developmental Medicine & Child Neurology 2016

 Age

IV Internal tibial torsion

Hip IR
In-toeing
Recurvatum
Pes varus
Equinus
Equinovarus

III Internal tibial torsion *

Hip IR
In-toeing *
Recurvatum
Pes varus
Equinus *
GMFCS

Equinovarus
II Internal tibial torsion *
Hip IR *
In-toeing *
Recurvatum
Pes varus
Equinus *
Equinovarus

I Internal tibial torsion *

Hip IR
In-toeing *
Recurvatum
Pes varus
Equinus *
Equinovarus
0 0.2 0.4 0.6 0.8 1 1.2 1.4
OR
*p<0.05

Prior surgery

IV Internal tibial torsion

Hip IR
In-toeing *
Recurvatum
Pes varus
Equinus *
Equinovarus *
III Internal tibial torsion
Hip IR *
In-toeing *
Recurvatum
Pes varus
GMFCS

Equinus *
Equinovarus

II Internal tibial torsion

Hip IR
In-toeing
Recurvatum
Pes varus
Equinus
Equinovarus

I Internal tibial torsion

Hip IR
In-toeing
Recurvatum
Pes varus
Equinus
Equinovarus
0 5 10 45
OR

Figure 2: Adjusted (multivariate) odds ratios (OR) (95% confidence interval) of gait deviations associated with equinus and in-toeing, changes with age
and prior surgery. GMFCS, Gross Motor Function Classification System; IR, internal rotation.

Prevalence of Gait Abnormalities in CP Susan A Rethlefsen et al. 5

GMFCS level I (including rotational malalignment, exter-
nal tibial torsion, pes valgus, excessive hip flexion, and
excessive knee flexion), level II (rotational malalignment,
pes valgus, out-toeing, and calcaneus), and level III (rota-
tional malalignment, external tibial torsion, pes valgus, cal-
caneus, out-toeing, and excessive hip flexion). The largest
impact of age on crouch and out-toeing was seen for rota-
tional malalignment (OR 1.32, 95% CI 1.15–1.51) and
external tibial torsion (OR 1.25, 95% CI 1.08–1.45) in
GMFCS level I. Odds of having excessive knee flexion
were significantly greater with age only in GMFCS level I
(OR 1.08, 95% CI 1.01–1.16). Odds of having calcaneus
were significantly increased with age in level II (OR 1.07,
95% CI 1.00–1.13) and level III (OR 1.15, 95% CI 1.07–
1.23) (Fig. 3, Table SIII). Age did not affect the odds of
having deviations associated with crouch and out-toeing in
GMFCS level IV.
Controlling for age and sex, the odds of having crouch
and out-toeing increased when prior surgery had been
done in GMFCS level I (out-toeing), level II (rotational
malalignment, excessive hip flexion, external rotation, out-
toeing), level III (out-toeing, calcaneus), and level IV (out-
toeing, calcaneus, rotational malalignment, and pes valgus).
The largest effects were seen in GMFCS level IV, with
increased odds of having calcaneus (OR 5.91, 95% CI
2.32–15.02), rotational malalignment (OR 5.24, 95% CI
1.45–19), and out-toeing (OR 7.14, 95% CI 2.69–18.94)
associated with prior surgery (Fig. 3, Table SV).
Other deviations
Controlling for age and sex, the odds of having stiff knee
gait increased significantly in all GMFCS levels when prior
surgery had been done. Odds of having stiff knee increased
the least in GMFCS level I (OR 2.37, 95% CI 1.21–4.64)
and the most in level IV (OR 5.31, 95% CI 1.41–20.06).
The prevalence of excessive hip adduction and anterior
and posterior pelvic tilt were unrelated to prior surgery
(Fig. 4, Table SV). Controlling for sex and prior surgery,
the odds of having excessive hip adduction, stiff knee, ante-
rior and posterior pelvic tilt were unchanged with increas-
ing age.
DISCUSSION
The current study represents the largest series to date eval-
uating the prevalence of gait deviations in patients with
CP, with more than 1000 patients in GMFCS levels I to
IV. The current study extends the important recent work
of Ounpuu et al.5 and provides more generalizable data.
The current study differs from that of Ounpuu et al. in
that: (1) the sample size of the current study is more than
three times greater; (2) participants in GMFCS level IV
are also included in this study; and (3) both unilaterally-
and bilaterally-involved children are included. Together
the studies confirm a higher prevalence (current study) and
greater severity (Ounpuu et al.) of gait deviations with
increasing GMFCS level (greater ambulatory disability). By
stratifying patients by GMFCS level, the current study
6 Developmental Medicine & Child Neurology 2016
improves on our previous report, and facilitates use of
these data by other healthcare providers.
The current study uses the GMFCS level, rather than
each patient’s limb distribution of CP, to group patients.
Cognitive, behavioral, and other factors can influence a
patient’s ambulatory abilities and, therefore, their GMFCS
level assignment. Gait deviations could be expected to be
better associated with physical characteristics such as limb
distribution. In the current study, 57% of participants in
GMFCS level I were classified as having hemiplegia, 51%
to 52% of participants in levels II and III had diplegia, and
59% of patients in level IV had quadriplegic CP, suggest-
ing some relationship between limb distribution and
GMFCS level, as has been found previously.7 The
GMFCS is a valid and reliable classification system, and
has good predictive value.4,8–10 As the GMFCS is now
used almost universally in the description of children with
CP, the data from the current study provide necessary
information for the clinician when considering common
gait problems in children with CP.
Overall, the most common gait deviation seen in the
current study was in-toeing (present in 70% of patients),
followed by excessive knee flexion (66%), stiff knee (63%),
and excessive hip flexion and adduction (57% each). These
results from an increased number of patients are similar to
our previous study,1 but with a few differences. Stiff knee
was the most prevalent deviation in the previous study, and
equinus was also more common than in the current study.
For participants in GMFCS levels I, II, and III, there
was evidence of progressive calcaneal crouch and out-toe-
ing associated with increasing age alone (increased odds of
calcaneus, pes valgus, excessive knee flexion, excessive hip
flexion, out-toeing, external tibial torsion, and rotational
malalignment; decreased odds of equinus, in-toeing, and
internal tibial torsion). This finding suggests that aggres-
sive treatments for equinus (such as tendo-Achilles length-
ening) should be avoided early in life if possible,
particularly in patients in GMFCS levels III and IV in
whom it might accelerate the progressive calcaneus that
occurs over time, potentially hastening functional decline.
Similarly, the likelihood of progressive out-toeing with age
should be taken into consideration when planning treat-
ments for in-toeing at a young age. We did not attempt to
correlate gait deviations prevalent in cases of prior surgery
with the specific surgical procedures performed, so no cau-
sal relationships can be inferred. However, the use of pre-
operative gait analysis data is known to alter treatment
decisions as well as the treatments performed.11–15 The
impact of gait analysis-directed surgery on the prevalence
of gait deviations may be different, and is an area for
future study.
Excessive knee flexion in stance was very common, being
present in more than 60% of participants in GMFCS levels
II to IV, and in more than 45% of those in level I. Odds
of having excessive knee flexion in stance increased with
increasing age in GMFCS levels I, II, and III, but this only
reached the level of statistical significance in participants in
 Age
IV Rotational malalignment
External tibial torsion
Hip flexion
Out-toeing
Excessive knee flexion
Pes valgus
Calcaneus

III Rotational malalignment *

External tibial torsion *
Hip flexion *
Out-toeing *
Excessive knee flexion
GMFCS

Pes valgus *
Calcaneus *

II Rotational malalignment *
External tibial torsion
Hip flexion
Out-toeing *
Excessive knee flexion
Pes valgus *
Calcaneus *

I Rotational malalignment *
External tibial torsion *
Hip flexion *
Out-toeing
Excessive knee flexion *
Pes valgus *
Calcaneus
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
OR
*p<0.05
Prior surgery

IV Rotational malalignment *
External tibial torsion
Hip flexion
Out-toeing *
Excessive knee flexion
Pes valgus *
Calcaneus *

III Rotational malalignment

External tibial torsion
Hip flexion
Out-toeing *
Excessive knee flexion
Pes valgus
GMFCS

Calcaneus *

II Rotational malalignment *
External tibial torsion *
Hip flexion *
Out-toeing *
Excessive knee flexion
Pes valgus
Calcaneus

I Rotational malalignment
External tibial torsion
Hip flexion
Out-toeing *
Excessive knee flexion
Pes valgus
Calcaneus
0 5 10 15 20 25 30 35
OR

Figure 3: Adjusted (multivariate) odds ratios (OR) (95% confidence interval) of gait deviations associated with crouch and out-toeing, changes with age
and prior surgery. GMFCS, Gross Motor Function Classification System.

Prevalence of Gait Abnormalities in CP Susan A Rethlefsen et al. 7

 Age

IV Posterior pelvic tilt

Anterior pelvic tilt
Hip adduction
Stiff knee

III Posterior pelvic tilt

Anterior pelvic tilt
Hip adduction
GMFCS

Stiff knee

II Posterior pelvic tilt

Anterior pelvic tilt
Hip adduction
Stiff knee

I Posterior pelvic tilt

Anterior pelvic tilt
Hip adduction
Stiff knee

0 0.2 0.4 0.6 0.8 1 1.2

OR
*p<0.05

Prior surgery

IV Posterior pelvic tilt

Anterior pelvic tilt
Hip adduction
Stiff knee *

III Posterior pelvic tilt

Anterior pelvic tilt
Hip adduction
Stiff knee *
GMFCS

II Posterior pelvic tilt

Anterior pelvic tilt
Hip adduction
Stiff knee *

I Posterior pelvic tilt

Anterior pelvic tilt
Hip adduction
Stiff knee *

0 5 10 15 20
OR

Figure 4: Adjusted (multivariate) odds ratios (OR) (95% confidence interval) of other gait deviations, changes with age and prior surgery. GMFCS, Gross
Motor Function Classification System.

8 Developmental Medicine & Child Neurology 2016

level I. This finding is consistent with our previous study,
which found a significant increase in odds of crouch with
increasing age only in participants with diplegia. It is also
in agreement with studies of the natural history of gait in
CP, which showed increased magnitude of crouch with age
in primarily independently ambulatory children with
CP.16–19 The current study is the first to examine a signifi-
cant number of patients in GMFCS levels III and IV. The
current study looked solely at presence or absence of
excessive knee flexion in stance, rather than the magnitude
of the problem. It is possible that, for these limited ambu-
lators, the magnitude of crouch increases with age but its
prevalence does not.
Combining the current study results with our previous
report, typical patients functioning in GMFCS level I can
be expected to have equinus, in-toeing, excessive knee flex-
ion, and stiff knee gait. Those classified as functioning in
GMFCS level II can be expected to have stiff knee gait,
excessive knee flexion, in-toeing, excessive hip flexion, and
equinus. Those functioning in GMFCS levels III and IV
can be expected to have the same problems as those in
GMFCS level II, with the addition of excessive hip adduc-
tion and pes valgus. Similar to the study by Ounpuu
et al.,5 there is overlap in the prevalence rates of gait prob-
lems among patients in the various GMFCS levels.
It is imperative that the clinician understands that there
is no universal combination of recommended surgeries
based solely on GMFCS level and patient age. Rather,
thorough consideration of patient-specific concerns, physi-
cal examination findings, and gait assessment are all essen-
tial for optimal surgical planning. Computerized gait
analysis testing allows a precise and objective evaluation
which optimizes such planning.13,20–22
The current study had additional limitations. Because all
of our patients had ambulatory ability and were seeking
orthopedic intervention to improve their ambulatory func-
tion, their gait problems may not reflect those of the gen-
eral CP population. Mildly involved patients who did not
seek or require intervention and non-ambulatory patients
were not represented. Gait problems were identified based
on clinical examination and kinematic data regardless of
their severity. Therefore, the problems identified may or
may not have impacted patient function. The gait problem
definitions were based on review of the literature as well as
our clinical judgement, and may differ from those used
elsewhere. We did not attempt to relate gait problems to
their possible causes, and some may be caused by multiple
factors. The kinematic model used in this study has
limitations, including modeling the foot as a single
segment rather than one with multiple joints. Therefore, it
is possible that some participants identified as having ‘cal-
caneus’ might have had midfoot breakdown. The kinematic
data were not normalized to gait speed, and walking
velocity was not controlled during gait testing. Therefore,
some deviations, such as stiff knee gait (partially defined as
delayed peak swing knee flexion) may have been related to
slow walking speed. Some patients used assistive devices
during gait analysis testing, which could have impacted
kinematic data and, therefore, prevalence of gait deviations
such as excessive hip flexion.
In summary, though the prevalence of specific gait devi-
ations varies according to GMFCS level, increased preva-
lence of calcaneal gait, out-toeing, and rotational
malalignment was seen with both age and prior surgery for
participants at all levels. This suggests that treatment deci-
sions for equinus and in-toeing should be made with cau-
tion in younger children, with consideration of the
potential for these problems to decrease naturally with age.
The study results will be useful to clinicians in their dis-
cussions with patients and families regarding the child’s
potential for developing various gait problems in the
future, based on GMFCS level. It will help clinicians to
tailor their treatment interventions for patients in GMFCS
levels I to IV to minimize the risk of developing gait prob-
lems which tend to become more prevalent as patients age
and/or undergo surgery.
A CK N O W L E D G E M E N T
The authors have stated that they had no interests which might
be perceived as posing a conflict or bias.
SUPPORTING INFORMATION
The following additional material may be found online:
Figure S1: Prevalence of gait abnormalities by GMFCS level
(color version).
Table SI: Definition of gait abnormalities based on three-
dimensional kinematic data and physical examination measures
Table SII: Prevalence of gait deviations by GMFCS level
Table SIII: Univariate analysis, OR for gait deviations with
increasing age, prior surgery, and increasing GMFCS level
Table SIV: Adjusted OR, changes with age (95% CI) adjusted
for sex and prior surgery
Table SV: Adjusted OR, changes with surgery (95% CI)
adjusted for sex and age

REFERENCES
1. Wren TA, Rethlefsen S, Kay RM. Prevalence of specific
gait abnormalities in children with cerebral palsy: influ-
ence of cerebral palsy subtype, age, and previous sur-
gery. J Pediatr Orthop 2005; 25: 79–83.
2. Rosenbaum PL, Palisano RJ, Bartlett DJ, Galuppi BE,
Russell DJ. Development of the Gross Motor Function
Classification System for cerebral palsy. Dev Med Child
Neurol 2008; 50: 249–53.
3. Palisano R, Rosenbaum P, Walter S, Russell D, Wood
E, Galuppi B. Development and reliability of a system
to classify gross motor function in children with cerebral
palsy. Dev Med Child Neurol 1997; 39: 214–23.
4. Palisano RJ, Rosenbaum P, Bartlett D, Livingston MH.
Content validity of the expanded and revised Gross
Motor Function Classification System. Dev Med Child
Neurol 2008; 50: 744–50.
5. Ounpuu S, Gorton G, Bagley A, et al. Variation in kine-
matic and spatiotemporal gait parameters by Gross

Prevalence of Gait Abnormalities in CP Susan A Rethlefsen et al. 9

 Motor Function Classification System level in children
and adolescents with cerebral palsy. Dev Med Child Neu-
rol 2015; 57: 955–62.
6. Monaghan K, Delahunt E, Caulfield B. Increasing the
number of gait trial recordings maximises intra-rater
reliability of the CODA motion analysis system. Gait
Posture 2007; 25: 303–15.
7. Gorter JW, Rosenbaum PL, Hanna SE, et al. Limb dis-
tribution, motor impairment, and functional classifica-
tion of cerebral palsy. Dev Med Child Neurol 2004; 46:
461–67.
8. Soo B, Howard JJ, Boyd RN, et al. Hip displacement in
cerebral palsy. J Bone Joint Surg Am 2006; 88: 121–29.
9. Rosenbaum PL, Walter SD, Hanna SE, et al. Prognosis
for gross motor function in cerebral palsy: creation of
motor development curves. JAMA 2002; 288: 1357–63.
10. Hanna SE, Bartlett DJ, Rivard LM, Russell DJ. Refer-
ence curves for the Gross Motor Function Measure: per-
centiles for clinical description and tracking over time
among children with cerebral palsy. Phys Ther 2008; 88:
596–607.
11. Kay RM, Dennis S, Rethlefsen S, Reynolds RA, Skaggs
DL, Tolo VT. The effect of preoperative gait analysis
10 Developmental Medicine & Child Neurology 2016
on orthopaedic decision making. Clin Orthop Relat Res
2000; 372: 217–22.
12. Kay RM, Dennis S, Rethlefsen S, Skaggs DL, Tolo VT.
Impact of postoperative gait analysis on orthopaedic
care. Clin Orthop Relat Res 2000; 372: 259–64.
13. DeLuca PA, Davis RB 3rd, Ounpuu S, Rose S, Sirkin R.
Alterations in surgical decision making in patients with
cerebral palsy based on three-dimensional gait analysis.
J Pediatr Orthop 1997; 17: 608–14.
14. Lofterod B, Terjesen T, Skaaret I, Huse AB, Jahnsen R.
Preoperative gait analysis has a substantial effect on
orthopedic decision making in children with cerebral
palsy: comparison between clinical evaluation and gait
analysis in 60 patients. Acta Orthop 2007; 78: 74–80.
15. Lofterod B, Terjesen T. Results of treatment when
orthopaedic surgeons follow gait-analysis recommenda-
tions in children with CP. Dev Med Child Neurol 2008;
50: 503–09.
16. Johnson DC, Damiano DL, Abel MF. The evolution of
gait in childhood and adolescent cerebral palsy. J Pediatr
Orthop 1997; 17: 392–96.
17. Rose GE, Lightbody KA, Ferguson RG, Walsh JC,
Robb JE. Natural history of flexed knee gait in diple-
gic cerebral palsy evaluated by gait analysis in children
who have not had surgery. Gait Posture 2010; 31:
351–54.
18. Gough M, Eve LC, Robinson RO, Shortland AP. Short-
term outcome of multilevel surgical intervention in spas-
tic diplegic cerebral palsy compared with the natural his-
tory. Dev Med Child Neurol 2004; 46: 91–97.
19. Bell KJ, Ounpuu S, DeLuca PA, Romness MJ. Natural
progression of gait in children with cerebral palsy. J
Pediatr Orthop 2002; 22: 677–82.
20. Kay RM, Dennis S, Rethlefsen S, Reynolds RA, Skaggs
DL, Tolo VT. The effect of preoperative gait analysis
on orthopaedic decision making. Clin Orthop Relat Res
2000; 372: 217–22.
21. Lofterod B, Terjesen T. Results of treatment when
orthopaedic surgeons follow gait-analysis recommenda-
tions in children with CP. Dev Med Child Neurol 2008;
50: 503–09.
22. Wren TA, Lening C, Rethlefsen SA, Kay RM. Impact
of gait analysis on correction of excessive hip internal
rotation in ambulatory children with cerebral palsy: a
randomized controlled trial. Dev Med Child Neurol 2013;
55: 919–25.