G Model

GAIPOS-3267; No. of Pages 4

Gait & Posture xxx (2011) xxx–xxx

Contents lists available at ScienceDirect

Gait & Posture

journal homepage: www.elsevier.com/locate/gaitpost

Evaluation of a single leg stance balance test in children

Thomas Zumbrunn b, Bruce A. MacWilliams a,b,c,*, Barbara A. Johnson a
a
Shriners Hospitals for Children, Salt Lake City, UT, USA
b
Department of Bioengineering, University of Utah, Salt Lake City, UT, USA
c
Department of Orthopaedic Surgery, University of Utah, Salt Lake City, UT, USA

A R T I C L E I N F O A B S T R A C T

Article history: Balance is a major determinate of gait. In high functioning individuals without significant vestibular or
Received 23 March 2010 vision impairments, a ceiling effect may be present when using a double limb support protocol to assess
Received in revised form 24 March 2011 balance function. For these populations, a single leg stance protocol may be more suitable. 47 typically
Accepted 5 April 2011
developing (TD) subjects and 10 patients with CEV performed a single leg stance test on a force plate. The
center of pressure (COP) was determined and several COP derived variables were calculated. Included
Keywords: measurements were: standard deviation, maximum excursion, area, average radial displacement, path
Balance
velocity and frequency of the COP. Directional components of suitable variables were used to analyze
Center of pressure
anterior/posterior and medial/lateral contributions. Correlations with age of TD subjects indicated that
Force plate
Children all balance variables except frequency were significantly correlated. Most parameters were highly inter-
Equinovarus correlated. Age adjusted COP balance variables also correlated to the Bruininks-Oseretsky balance
Club foot subtest. Highest correlations were determined by the maximum excursion and velocity of the COP in the
anterior/posterior direction. Statistical comparisons between the CEV group and a 4–6 TD group
indicated significant differences between groups for most COP balance parameters. These results
indicated that a single limb balance assessment may be a useful assessment for determining balance
impairments in higher functioning children with orthopedic impairments.
ß 2011 Elsevier B.V. All rights reserved.

1. Introduction limb support during gait. Additionally high functioning individuals

without vestibular or visual impairments may not be sufficiently
Postural stability or balance is the ability to maintain the body challenged using double support tests. In this case double limb
in equilibrium by keeping the projected center of mass (COM) support protocols may fail to detect longitudinal changes or
within the limits of the base of support [1]. According to Gage, differences compared to controls.
stability in stance is the first prerequisite for normal gait [2]. There There are few reports of single limb balance investigations. The
is a repeated and rhythmic alternation of double and single leg Romberg Test is a timed single leg stance protocol, but has not been
support during walking with periods of single support necessary to used with force plate data [7–9]. Reports of single limb balance using
allow swing phase mechanics of the opposite limb; hence the force plate measures have been limited to investigations of ankle
ability to control the COM during single leg stance is important for instability in adults [10–14] and an analysis of balance differences
independent walking [3]. between genders in nine to eleven year old children [15].
Humans use multiple sensory systems to maintain an upright Congenital talipes equinovarus (CEV), also referred to as
posture including the vestibular system, the visual system and the clubfoot, is one of the most common musculoskeletal birth defects
somatosensory system. Most previous work has focused on [16]. It occurs in one of 1000 live births and is considered idiopathic
children with neuro-motor impairments and used double leg since it occurs mostly as an isolated birth defect [17]. CEV consists
stance for balance testing [4–6,3]. These protocols usually include of bony deformities, soft tissue contractures and muscular
eyes open and eyes closed test-variations. For subjects without weakness. The foot is positioned in supination, adduction and
neurological conditions, these protocols may not test the ability of varus when standing. Compared to normal feet, the smaller
the musculoskeletal system to provide functional balance in single support area, muscular disbalance and joint instability may impair
balance in patients with CEV [16,18].
The first aim of this study was to describe a single leg balance
test on a force plate to quantify typically developing children’s
* Corresponding author at: Motion Analysis Laboratory, Shriners Hospitals for
ability to maintain single limb support. Developmental motor tests
Children-Salt Lake City, Fairfax Rd. @ Virginia St., Salt Lake City, UT 84103, USA.
Tel.: +1 801 536 3800; fax: +1 801 536 3782. for children have documented increased ability to balance in single
E-mail address: [email protected] (B.A. MacWilliams). limb support with maturation [19]. We hypothesized that the

0966-6362/$ – see front matter ß 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.gaitpost.2011.04.005

Please cite this article in press as: Zumbrunn T, et al. Evaluation of a single leg stance balance test in children. Gait Posture (2011),
doi:10.1016/j.gaitpost.2011.04.005
G Model
GAIPOS-3267; No. of Pages 4

2 T. Zumbrunn et al. / Gait & Posture xxx (2011) xxx–xxx

Table 1
COP force plate variables used to evaluate single limb balance [6,14]. In our calculations x represents the anterior/posterior (A/P) and y represents the medial/lateral (M/L)
components. All data are computed over a three second period.

Measure Definition

COPsd Standard deviation of the COP, resultant or in a given direction (x, y) expressed in mm.
COPmax The absolute maximum distance of the COP excursion in a given direction expressed in mm.
COParea A rectangular area defined by the maximum excursion of the COP in anterior/posterior and medial/lateral
directions Area = (xmax  xmin)  (ymax  ymin) expressed in mm2. qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Average radial displacement (ARD) Average of the resultant displacements with respect to the mean COP: ARD ¼ ðx  x̄Þ2 þ ðy  ȳÞ2 , expressed
in mm.
Path velocity Average distance traveled per second of the COP (V) expressed in mm/s.
COPvel Directional components of path velocity V expressed in mm/s.
Average radial frequency (ARF) COP frequency determined by average radial displacement and the path velocity ARF = V/2pARD expressed in cycles/second.

center of pressure could be used to quantify balance and that the analyzed for the CEV group. For children with bilateral limb involvement in the CEV
group, data from one side was randomly selected using a computer generated list.
data would show improvement in single limb support of typically
Most control subjects were tested on both legs and a random limb was selected for
developing subjects with increasing age. The second aim of the inclusion using a computer generated list.
study was to compare a cohort of CEV patients to a similarly aged
typically developing cohort. We hypothesized that the single limb 2.5. BOT-2 balance subtest
balance protocol would be sensitive enough to determine balance
The balance subtest of the Bruininks-Oseretsky Test of Motor Proficiency
impairments in the CEV group. The third aim of the study was to
instrument (second edition, BOT-2) was administered to the CEV subjects [20]. The
explore the relationship between single limb balance center of BOT-2 is an individually administered measure of fine and gross motor skills of
pressure measures and a standardized balance test. We hypothe- children 4–21 years of age. The test consists of eight subtests including one for
sized that some significant correlations would exist between these balance assessment. The age/gender matched mean scores reported in the BOT-2 for
measures. the balance subtest were used for comparison to CEV subjects.

2. Methods 2.6. Statistics

2.1. Subjects Correlations of all balance parameters with age were first assessed within the
entire TD cohort (SPSS, Chicago, IL, USA). Parameters with significant age
Forty-seven typically developing (TD) subjects (age 4–18) and 10 children with
correlations were further described by clinical age groupings: 4–7, 8–12 and
congenital equinovarus (CEV) who were previously treated with Ponsetti casting
13–18 (Table 2) to correspond to previously established criteria [20]. The CEV
(age 4–6), were evaluated in separate IRB approved studies. After attaining
group was compared to a 4–6 TD group using Mann–Whitney U-tests.
informed consent, the subjects were tested in the motion analysis laboratory by one
Significance level was set at a = 0.05 for all tests. Correlations between all
of two experienced physical therapists. The assessment included balance testing
force plate parameters and the BOT-2 score were examined after adjusting for
using a force plate and a standardized balance assessment (CEV group only). Two
age. Both TD subjects and CEV patients were included in the correlation to
additional children in each group were enrolled but could not complete the balance
provide a spectrum of coordination.
task.

2.2. Balance protocol 3. Results

Data were recorded at 1000 Hz from a force plate (AMTI OR-6, Watertown, MA,
USA) integrated in the laboratory floor. Subjects were positioned to face along a Plots of all data and correlations between all variables from the
laboratory based coordinate system, which was coincident with the embedded TD group are given in an electronic appendix. Correlation analysis
force plate coordinate system. They were instructed to stand in the center of the (Table 3) indicated that all balance parameters considered from
force plate and on the examiner’s command to stand on one foot for five seconds.
typically developing children were significantly correlated with
Data collection began during double limb stance and continued until the subject
stood on one leg for the count of five. Five trials were collected standing on each leg. age except ARF. ARF was therefore not considered for further
Criterion for a valid trial was standing on one foot for three continuous seconds analysis. All parameters show negative correlations with increas-
without having the contralateral foot touch the ground. The three second interval ing age indicating improved balance. Age was most highly
was chosen where the center of pressure (COP) excursion was smallest by visual correlated (jrj > 0.5) with ARD, COPmax A/P, COPsd Res, COPsd
inspection of the force plate vector data. This typically occurred toward the end of
the 5 s as the subjects found balance equilibrium after initial counterbalance. This
A/P and COPvel A/P. The means and standard deviations of all
was also demonstrated by Rival et al. using a similar protocol [5]. remaining COP variables are listed in Table 4. Inter-correlations
between most force plate parameters were high indicating they are
2.3. Center of pressure measurements similar measures (Table A.1).
The comparison between the CEV subjects and the similarly
Several balance variables were calculated to enable comparison to previous
reports and assess sensitivities. COP standard deviation (COPsd), maximum
aged TD group showed significant between group differences both
excursion (COPmax), and velocity (COPvel) were each computed with anterior/ for all force plate parameters (except COPsd M/L) and for BOT-II
posterior (A/P) and medial/lateral (M/L) directional components. Additionally, the
resultant (summed A/P and M/L components) COP standard deviation (COPsd Res),
area (COParea), average radial displacement (ARD), path velocity and average radial Table 2
frequency (ARF) were computed using previously defined methods (Table 1) Typically developing (TD) age groups and congenital equinovarus (CEV) subjects
[6,14,3]. with gender, age and number of limbs tested. Means  standard deviations are listed
for age.

2.4. Data processing Group Subjects (M/F) Age [Months]

Force plate data were input to a custom program to compute all COP based Clinical age groups
variables. A zero-lag fourth-order Butterworth filter with a 10 Hz cut-off frequency TD (4–7) 15 (9/6) 74  13
was applied. For determining group means and SD, the best trial performed by each TD (8–12) 19 (8/11) 119  15
subject for each leg was selected using minimum COPsd Res as criterion. Many trials TD (13–18) 13 (7/6) 178  16
either did not meet the 3 s requirement or had notable balance excursions. These CEV and age matched controls
trials were excluded to prevent data skewing; more consistent data was obtained by TD (4–6) 11 (7/4) 68  10
taking the best trial rather than the average. Only the affected limb data was CEV (4–6) 10 (7/3) 63  7

Please cite this article in press as: Zumbrunn T, et al. Evaluation of a single leg stance balance test in children. Gait Posture (2011),
doi:10.1016/j.gaitpost.2011.04.005
G Model
GAIPOS-3267; No. of Pages 4

T. Zumbrunn et al. / Gait & Posture xxx (2011) xxx–xxx 3

Table 3
Correlations of all calculated force plate variables with age from typically developing children aged 4–18 (N = 47 limbs). Pearson’s correlation coefficient (r) and two-tailed
significance (p).

Variable COPsd A/P COPsd M/L COPsd Res COPmax A/P COPmax M/L COParea ARD Path Velocity COPvel A/P COPvel M/L ARF

r 0.525 0.478 0.558 0.561 0.419 0.491 0.564 0.484 0.505 0.386 0.063
p <0.001 0.001 <0.001 <0.001 0.003 <0.001 <0.001 <0.001 <0.001 0.007 0.673

Table 4 4. Discussion
Descriptive statistics of single leg stance values of all force plate variables having
significant correlations with age (mean  SD; A/P, anterior/posterior; M/L. medial/
lateral; Rd, radial; Res, resultant) and BOT-II scores for all typically developing (TD) age Understanding the role of balance in gait and other activities
groups. may assist with diagnosis and treatment planning. This study
aimed to evaluate a single limb stance balance protocol targeted to
Parameter TD (4–7) TD (8–12) TD (13–18)
(n = 15) (n = 19) (n = 13)
measure musculoskeletal components of balance critical to
movement. The study also evaluated different variables used to
COPsd A/P [mm] 8.1  3.7 5.5  1.8 4.7  1.6
express balance to establish which were the most sensitive for this
COPsd M/L [mm] 6.3  2.9 5.4  1.2 3.8  0.8
COPsd Res [mm] 10.5  4.3 7.7  1.9 6.1  1.5 protocol, based on the ability to discriminate between groups.
COPmax A/P [mm] 35.2  13.3 23.1  7.2 19.2  6.6 The first hypothesis of this study, that center of pressure
COPmax M/L [mm] 23.9  10.2 22.3  5.8 15.9  3.8 parameters measured during single limb balance could be used to
COParea [mm2] 929  677 536.0  267 315  158
quantify balance and that the data would show improvement in
ARD [mm] 9.4  3.9 6.9  1.7 5.4  1.3
Path Velocity [mm/s] 67.5  33.8 54.7  17.4 37.4  12.8
single limb support of TD subjects with increasing age was
COPvel A/P [mm/s] 49.3  24.9 35.1  13.3 25.9  10.9 supported by the results. All but one variable studied here was
COPvel M/L [mm/s] 36.9  21.1 34.3  11.9 21.7  6.9 significantly correlated with age. Older subjects in the TD group
BOT-II Score 27.6  3.4 32.9  0.7 33.3  0.6 exhibited smaller variance indicative of improvement in balance
abilities with maturation.
The second hypothesis, that a single limb balance protocol
Table 5 would be sensitive enough to determine balance impairments in a
Single leg stance values of all force plate variables having significant correlations
cohort of children with CEV was also supported by the data.
with age (mean  SD; A/P, anterior/posterior; M/L, medial/lateral; Rd, radial; Res,
resultant) of 4–6 year old typically developing (TD) children and the 4–6 year old CEV Statistically significant balance differences were determined
patients. Age and BOT-II scores are also shown. The two groups are compared using between the CEV group and the similarly aged TD group in all
one-tailed Mann–Whitney U-tests. analyzed balance variables except COPsd M/L. Children with CEV
Parameter CEV TD (4–6 years) Mann–Whitney,
also scored significantly lower than age matched normative data in
(n = 10) (n = 11) p the BOT-2 balance subtest (p < 0.001), indicating delays in balance.
Many of the higher level balance tasks in the BOT-2 include single
Age (months) 63.4  6.9 68.4  10.0 0.217*
COPsd A/P [mm] 14.4  5.4 8.8  4.0 0.007 limb support. The force plate parameters indicated greater levels of
COPsd M/L [mm] 7.7  1.8 6.8  2.8 0.287 significance in the anterior/posterior direction compared to
COPsd Res [mm] 16.4  5.3 11.3  4.3 0.006 medial/lateral components (Table 5).
COPmax A/P [mm] 61.7  25.6 37.6  14.4 0.008
The third hypothesis, that significant correlations would exist
COPmax M/L [mm] 31.0  9.0 26.2  9.9 0.080
COParea [mm2] 2021  1227 1063  707 0.018 between center of pressure measurements and a pediatric balance
ARD [mm] 14.1  4.3 10.1  3.9 0.010 assessment was also supported by the data. All COP variables had
Path Velocity [mm/s] 110.2  36.0 72.3  31.9 0.006 negative relationships with BOT-2 scores, indicating that greater
COPvel A/P [mm/s] 89.0  35.5 52.3  21.7 0.005 coordination skills lead to improved balance (Table 6). Anterior/
COPvel M/L [mm/s] 49.6  13.6 40.5  22.5 0.023
posterior components were more highly correlated than medial/
BOT-II score 20.2  4.4 26.4  3.1 <0.001
lateral components. Only bi-directional or anterior/posterior
*
Two-tailed test for age comparison.
directed variables were significant. Further research is necessary
to establish the relationship between factors influencing COP
measures in single limb support, such as strength, or fine or gross
scores (Table 5). Comparison of the ages between groups motor proficiency.
supported the assumption that age differences did not significantly A tertiary aim of this study was to determine from a myriad
affect the comparisons. variable expressions all derived from the same raw center of
Age adjusted correlations of all force plate parameters with pressure data, which variables best measured balance for this
BOT-2 balance scores demonstrated significant (p < 0.05) correla- protocol. One variable, average radial frequency, did not correlate
tions only between bi-directional and anterior/posterior variables with age and was therefore not further considered. Most other
(Table 6). All medial lateral variables were not significantly variables were highly inter-correlated. Not surprisingly, distance
correlated. The BOT-2 balance scores were most highly correlated measures were less highly correlated with velocity measures. The
(jrj > 0.4) with COPvel A/P, COPsd A/P and COPmax A/P. most highly correlated (jrj > 0.5) force plate parameters to age

Table 6
Pearson’s correlation coefficient values and significance for partial correlations controlled for age between the force plate parameters and the BOT-2 score for all subjects
(n = 57).

Variable COPsd A/P COPsd M/L COPsd Res COPmax A/P COPmax M/L COParea ARD Path Velocity COPvel A/P COPvel M/L

r 0.427 0.065 0.365 0.422 0.111 0.328 0.352 0.385 0.462 0.164
p 0.001 0.635 0.006 0.001 0.416 0.014 0.008 0.003 <0.001 0.226

Please cite this article in press as: Zumbrunn T, et al. Evaluation of a single leg stance balance test in children. Gait Posture (2011),
doi:10.1016/j.gaitpost.2011.04.005
G Model
GAIPOS-3267; No. of Pages 4

4 T. Zumbrunn et al. / Gait & Posture xxx (2011) xxx–xxx

were ARD, COPmax A/P, COPsd Res, COPsd A/P and COPvel A/P. All Conflict of interest statement
are bidirectional or anterior/posteriorly directed. Although all but
one force plate parameter were significantly different in compar- The authors of this manuscript have no financial or personal
isons between similarly aged TD children and CEV patients, relationships with any peoples or organizations which could
bidirectional and A/P directed variables were generally more inappropriately influence this work.
highly significant. Finally in correlations across all subjects with
adjustments for age, only bidirectional and A/P directed variables
were significant, with COPvel A/P, COPsd A/P and COPmax A/P Appendix A. Supplementary data
demonstrating the highest correlations (jrj > 0.4). Thus both
COPvel A/P and COPmax A/P were among the highest correlated
Supplementary data associated with this article can be found, in
variables both with age and across the age adjusted population
the online version, at doi:10.1016/j.gaitpost.2011.04.005.
examined here. Additionally, these two variables were not as
highly inter-correlated (r = 0.8) as most other combinations. This
suggests that these two variables may be sufficient in reporting the References
results of this test.
[1] Shumway-Cook A, Woollacott MH. Motor control: theory and practical appli-
Studies using double limb support protocols report better cation. Baltimore, MD: Lipincott Williams & Wilkins; 2001.
balance values than our single limb protocol [6,3]. This is [2] Gage JR, DeLuca PA, Renshaw TS. Gait analysis: principle and applications with
understandable considering stance on both limbs is more centered emphasis on its use in cerebral palsy. Instr Course Lect 1996;45:491–507.
[3] Wolff DR, Rose J, Jones VK, Bloch DA, Oehlert JW, Gamble JG. Postural balance
and stable. Compared to previous results from single limb measurements for children and adolescents. J Orthop Res 1998;16:271–5.
protocols, similarities are evident regarding the reported values [4] Hertel J, Olmsted-Kramer LC, Challis JH. Time-to-boundary measures of pos-
of the directional variables; however, some values reported here in tural control during single leg quiet standing. J Appl Biomech 2006;22:67–73.
[5] Rival C, Ceyte H, Olivier I. Developmental changes of static standing balance in
children indicate better balance than those reported for typically children. Neurosci Lett 2005;376:133–6.
developed adults [12,14]. [6] Rose J, Wolff DR, Jones VK, Bloch DA, Oehlert JW, Gamble JG. Postural balance
There are several limitations in the present study. A double limb in children with cerebral palsy. Dev Med Child Neurol 2002;44:58–63.
[7] Balogun JA, Ajayi LO, Alawale F. Determinants of single limb stance balance
balance protocol was not tested in these subjects, so a direct
performance. Afr J Med Med Sci 1997;26:153–7.
comparison between protocols to assess relative sensitivity was [8] Freeman MA. Instability of the foot after injuries to the lateral ligament of the
not possible. BOT-2 tests scores were not acquired for the TD ankle. J Bone Joint Surg Br 1965;47:669–77.
[9] Freeman MA, Dean MR, Hanham IW. The etiology and prevention of functional
subjects used in the balance testing protocol, instead published
instability of the foot. J Bone Joint Surg Br 1965;47:678–85.
norms were used for comparison to the CEV group. Only one ‘‘best’’ [10] Brown CN, Mynark R. Balance deficits in recreational athletes with chronic
balance trial was used for analysis, thus results may not be ankle instability. J Athl Train 2007;42:367–73.
indicative of an individual’s typical balance, but rather more [11] Cornwall MW, Murrell P. Postural sway following inversion sprain of the ankle.
J Am Podiatr Med Assoc 1991;81:243–7.
indicative of their highest capability. No normalization for foot size [12] Ross SE, Arnold BL, Blackburn JT, Brown CN, Guskiewicz KM. Enhanced balance
was used in this study as this has not been done in previous associated with coordination training with stochastic resonance stimulation
literature and is beyond the scope of the current study. As smaller in subjects with functional ankle instability: an experimental trial. J Neuroeng
Rehabil 2007;4:47.
children had larger variable magnitudes, normalizing to foot size [13] Ross SE, Guskiewicz KM. Examination of static and dynamic postural stability
would only serve to increase the significant relationships found in individuals with functionally stable and unstable ankles. Clin J Sport Med
here. 2004;14:332–8.
[14] Ross SE, Guskiewicz KM, Gross MT, Yu B. Balance measures for discriminating
between functionally unstable and stable ankles. Med Sci Sports Exerc
5. Conclusion 2009;41:399–407.
[15] Lee AJ, Lin WH. The influence of gender and somatotype on single-leg upright
standing postural stability in children. J Appl Biomech 2007;23:173–9.
Using a single limb balance protocol, COP measures were able to [16] Dobbs MB, Gurnett CA. Update on clubfoot: etiology and treatment. Clin
differentiate balance performance between age groups and Orthop Relat Res 2009;467:1146–53.
differentiate between patients with CEV and a similarly aged [17] Wynne-Davies R. Family studies and the cause of congenital club foot. Talipes
equinovarus, talipes calcaneo-valgus and metatarsus varus. J Bone Joint Surg
typically developing cohort. Thus COP measures of single limb
Br 1964;46:445–63.
balance may be a useful assessment for determining balance [18] Miedzybrodzka Z. Congenital talipes equinovarus (clubfoot): a disorder of the
impairments in higher functioning children with orthopedic foot but not the hand. J Anat 2003;202:37–42.
impairments. Of the variables tested, the best correlations between [19] Folio MR, Fewell RR. Peabody developmental motor scales: examininer’s
manual. Austin, TX: PRO-ED, Inc; 2000.
both balance differences with age and with pathology were [20] Bruininks RH, Bruininks BD. The Bruininks-Oseretsky test of motor proficien-
COPmax A/P and COPvel A/P. cy. Circle Pines, MN: AGS Publishing; 2005.

Please cite this article in press as: Zumbrunn T, et al. Evaluation of a single leg stance balance test in children. Gait Posture (2011),
doi:10.1016/j.gaitpost.2011.04.005