JCLB-03739; No of Pages 7

Clinical Biomechanics xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Clinical Biomechanics
journal homepage: www.elsevier.com/locate/clinbiomech

Frontal and transverse plane hip kinematics and gluteus maximus

recruitment correlate with frontal plane knee kinematics during
single-leg squat tests in women
John H. Hollman ⁎, Christy M. Galardi, I-Hsuan Lin, Brandon C. Voth, Crystal L. Whitmarsh
Program in Physical Therapy, the Department of Physical Medicine & Rehabilitation and Sports Medicine Center, Mayo Clinic, Rochester, MN, USA

a r t i c l e i n f o a b s t r a c t

Article history: Background: Hip muscle dysfunction may be associated with knee valgus that contributes to problems like
Received 2 August 2013 patellofemoral pain syndrome. The purpose of this study was to (1) compare knee and hip kinematics and hip
Accepted 30 December 2013 muscle strength and recruitment between “good” and “poor” performers on a single-leg squat test developed
to assess hip muscle dysfunction and (2) examine relationships between hip muscle strength, recruitment and
Keywords:
frontal plane knee kinematics to see which variables correlated with knee valgus during the test.
Lower extremity
Methods: Forty-one active women classified via visual rating as “good” or “poor” performers on the test partici-
Hip
Knee
pated. Participants completed 5-repetition single-leg squat tests. Isometric hip extension and abduction strength,
Biomechanics gluteus maximus and gluteus medius recruitment, and 3-dimensional hip and knee kinematics during the test
Electromyography were compared between groups and examined for their association with frontal plane knee motion.
Muscle strength Findings: “Poor” performers completed the test with more hip adduction (mean difference = 7.6°) and flexion
(mean difference = 6.3°) than “good” performers. No differences in knee kinematics, hip strength or hip muscle
recruitment occurred. However, the secondary findings indicated that increased medial hip rotation (partial
r = 0.94) and adduction (partial r = 0.42) and decreased gluteus maximus recruitment (partial r = 0.35) cor-
related with increased knee valgus.
Interpretation: Whereas hip muscle function and knee kinematics did not differ between groups as we'd
hypothesized, frontal plane knee motion correlated with transverse and frontal plane hip motions and with
gluteus maximus recruitment. Gluteus maximus recruitment may modulate frontal plane knee kinematics
during single-leg squats.
© 2014 Elsevier Ltd. All rights reserved.

1. Introduction (Finnoff et al., 2011; Fredericson et al., 2000; Fulkerson, 2002;

The prevalence of patellofemoral pain syndrome, characterized by
peri- or retropatellar pain, in active adolescent and young adult
women is approximately 15% (Boling et al., 2010; Myer et al., 2010).
The condition is commonly aggravated by squatting, kneeling, travers-
ing stairs and running and therefore affects one's ability to perform ath-
letic activities (Dixit et al., 2007). Recent evidence suggests hip muscle
function is impaired in individuals with patellofemoral pain syndrome
and may be a risk factor for its development. Prins and Van Der Wurff
(2009) reported that hip abduction, external rotation and extension
strength are all diminished in women with patellofemoral pain
syndrome compared with asymptomatic individuals. Impaired hip
abduction and external rotation strength have also been identified as
risk factors for the condition in adolescent female runners. If hip abduc-
tion and external rotation strength are impaired, excessive adduction
and medial rotation during weightbearing activities may be induced
⁎ Corresponding author at: Mayo Clinic, Program in Physical Therapy, Siebens 11, 200
First Street SW, Rochester, MN 55905, USA.
E-mail address: hollman.john@mayo.edu (J.H. Hollman).
Fulkerson and Arendt, 2000; Ireland et al., 2003), which concomitantly
increases knee valgus that is theorized to contribute to patellofemoral
pain syndrome (Powers, 2003).
Impaired neuromuscular control may also contribute to movements
that trigger patellofemoral pain syndrome. Neuromuscular control–
defined as the maintenance of functional joint stability through subcon-
scious muscle recruitment in response to joint movements and loading
conditions (Riemann and Lephart, 2002)–is reflected by electromyo-
grams (EMG) that quantify neural input to muscles. Gluteus medius re-
cruitment, for example, is delayed in women with patellofemoral pain
syndrome during stair-stepping tasks compared with asymptomatic in-
dividuals (Brindle et al., 2003; Cowan et al., 2009). Moreover, peak glu-
teus maximus recruitment correlates negatively with knee valgus
during single leg squats (Hollman et al., 2009). These relationships
make sense because the gluteus medius abducts and the gluteus
maximus extends and laterally rotates the hip. Recruiting these muscles
during weightbearing tasks may limit hip adduction or medial rotation,
thereby limiting knee valgus, implying that altered neuromuscular con-
trol at the hip may influence hip and knee kinematics and contribute to
patellofemoral pain syndrome.

0268-0033/$ – see front matter © 2014 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.clinbiomech.2013.12.017

Please cite this article as: Hollman, J.H., et al., Frontal and transverse plane hip kinematics and gluteus maximus recruitment correlate with frontal
plane knee kinematics during single-leg squat test..., Clin. Biomech. (2014), http://dx.doi.org/10.1016/j.clinbiomech.2013.12.017
2 J.H. Hollman et al. / Clinical Biomechanics xxx (2014) xxx–xxx

To understand how hip muscle dysfunction may be associated with testing. Participants were screened for eligibility with a knee pain histo-
impaired movements that contribute to patellofemoral pain syndrome, ry questionnaire; with Lachman, posterior drawer, valgus and varus
Crossley et al. (2011) developed a rating system for assessing single-leg stress tests for ligament stability; a patellar apprehension test for
squat performance in healthy adults based on balance and perturba- patellofemoral instability; and palpation for joint line tenderness indic-
tions as well as the trunk, pelvis, hip and knee posture (Table 1). Partic- ative of meniscal injury. A single investigator performed the screening
ipants classified as “poor” performers with hip dysfunction had delayed exam. All participants provided written informed consent. The study
gluteus medius onset and weaker hip abduction strength than those was approved by the Mayo Foundation Institutional Review Board.
rated as “good” performers (Crossley et al., 2011). The investigators Since estimates of hip extension strength and variance are more
did not examine the magnitude of gluteus medius recruitment during readily available in the literature than are estimates of hip and knee
the test. Furthermore, despite the potential influences of gluteus kinematics or hip muscle recruitment during single-leg squats, we con-
maximus function on hip and knee kinematics as described above, ducted a power analysis based on previously published hip strength
the investigators did not examine hip extension strength or gluteus data. Hip extension strength in healthy, adult women is approximately
maximus recruitment, nor did they examine 3-dimensional hip and 43% body weight (% BW) (Hollman et al., 2012). Detecting a 10% BW
knee kinematics to validate their rating system. The purposes of this difference in hip extension strength, assuming a standard deviation of
study, therefore, were twofold. First, we compared hip and knee 11% BW within groups, at a statistical power of 0.80 required a sample
kinematics and hip muscle function–including hip extension and size of 42 participants (21 per group).
abduction strength and magnitudes of gluteus maximus and gluteus
medius recruitment–between “good” and “poor” performers on the
2.2. Instrumentation
single-leg squat test. Second, we examined the relationships between
hip strength, gluteus maximus and gluteus medius recruitment and
Data collected included participants' physical activity levels, hip ex-
hip and knee kinematics to see which variables most strongly predicted
tension and abduction muscle strength, gluteus maximus and gluteus
the magnitude of knee valgus during the test. We hypothesized that
medius recruitment and hip and knee kinematics during single-leg
knee valgus, hip adduction and medial hip rotation would be greater–
squats. Since physical activity may influence dynamic performance,
and hip extensor strength and gluteus maximus recruitment would be
participants completed a 7-question form of the International Physical
lower–in women who demonstrated poor performance during the
Activity Questionnaire (Craig et al., 2003). Performance of single-leg
single-leg squat test. Third, we hypothesized that reduced hip extension
squats during an initial screening exam was recorded with a Kodak
strength and gluteus maximus recruitment would be associated with
Zi8 digital video camera (Eastman Kodak Company, Rochester, New
increased knee valgus.
York, USA). Hip extension and abduction strength were measured
with a MicroFET2 dynamometer (Hoggan Health Industries, Inc., West
2. Methods
Jordan, Utah, USA). EMGs were acquired at 1000 Hz through a 16-bit
NI-DAQ PCI-6220 analog-to-digital card (National Instruments Corpora-
2.1. Participants
tion, Austin, Texas, USA) with Bagnoli DE-3.1 double-differential bipolar
surface electrodes and a Bagnoli-16 amplifier (Delsys Inc., Boston,
Active, healthy active women 18 to 36 years old were recruited.
Massachusetts, USA) having a common mode rejection ratio of 92 dB
Inclusion criteria included the ability to walk, run, jump, and squat with-
at 60 Hz, input impedance N 1015 ohms, estimated noise ≤ 1.2 μV and
out knee pain. Exclusion criteria included history of knee ligament inju-
overall amplification of 100–10,000 V/V. Electrodes were constructed
ry, patella dislocation, knee pain within 6 months of testing, or any
of 10-mm × 1-mm silver bars, spaced 10-mm apart within preampli-
suspected pathology of the knee or hip that compromised safety during
fiers having a gain of 10 V/V. Kinematic data were acquired at 100 Hz
with a Vicon MX system and five MX20 + cameras (Vicon Motion
Table 1
Rating criteria for the single-leg squat test. Performance was operationally defined Systems, Oxford, United Kingdom). Vicon Nexus software was used to
as “Good” when participants performed the test over 5 repetitions with 3 or fewer quantify lower extremity kinematics.
departures from the Crossley et al. (2011) criteria for good performance; as “Fair” perfor-
mance with 4–7 departures from the criteria; and as “Poor” performance when participants
performed the test with 7 or more departures from the criteria. 2.3. Procedures
Criterion To be rated “Good”
Following a physical exam to assess eligibility for enrolling in the
Overall impression across 5 squats study, the single-leg squat test was demonstrated. Participants crossed
Ability to maintain balance Participant does not lose balance
Perturbations of the person Movement is performed smoothly
their arms in front of their chest and performed single-leg squats with
Depth of the squat Performed to at least 60° their dominant leg (preferred kicking leg) five times consecutively
of knee flexion from a 20-cm high platform at a rate of approximately 1 squat per 2 s,
Speed of the squat Performed at approximately 1 per 2 s as described by Crossley et al. (2011). Participants performed 3 practice
Trunk posture
squats to familiarize themselves with the task before their performance
Trunk/thoracic lateral deviation No trunk/thoracic lateral deviation or shift
or shift was formally tested. In contrast to methods used by Crossley et al.
Trunk/thoracic rotation No trunk/thoracic rotation (2011), who used standardized footwear, our participants performed
Trunk/thoracic lateral flexion No trunk/thoracic lateral flexion the test with bare feet to eliminate potential effects of footwear on
Trunk/thoracic forward flexion No trunk/thoracic forward flexion lower extremity mechanics. Test performance was recorded with the
Pelvis posture
Pelvic lateral deviation No pelvic lateral deviation
digital video camera placed on a tripod 3-m anterior to the participant
Pelvic rotation No pelvic rotation and downloaded onto a Dell Optiplex 755 desktop computer.
Pelvic tilt (take note of depth of squat) No pelvic tilt Four investigators independently reviewed the video footage and
Hip joint rated each participant's performance as “good,” “fair” or “poor” based
Hip adduction No hip adduction
on Crossley et al.'s (2011) criteria (Table 1). Agreement by three or
Hip medial rotation No hip medial rotation
Knee joint more raters was required to categorize performance as “good” or
Apparent knee valgus No apparent knee valgus “poor.” On analysis of the investigators' ratings, overall agreement was
Knee position relative to foot position Center of knee remains 70% and the kappa coefficient for multiple raters was equal to 0.55,
over center of foot slightly lower than kappa coefficients from 0.60 to 0.80 reported by
Note: adapted from Crossley et al. (2011) with permission. Crossley et al. (2011).

Please cite this article as: Hollman, J.H., et al., Frontal and transverse plane hip kinematics and gluteus maximus recruitment correlate with frontal
plane knee kinematics during single-leg squat test..., Clin. Biomech. (2014), http://dx.doi.org/10.1016/j.clinbiomech.2013.12.017
 J.H. Hollman et al. / Clinical Biomechanics xxx (2014) xxx–xxx 3

Participants classified as “good” and “poor” performers returned

Fig. 1. Markers were placed on anatomic landmarks in accordance with Vicon's Plug-in
Gait marker set at the posterior and anterior superior iliac spines, the lateral midline
and lateral epicondyle of the thighs, the lateral midline and lateral malleoli of the shanks,
posterior aspect of the calcanei and dorsum of the 2nd metatarsophalangeal joints.
within one week for repeat testing. Since there is evidence functional
lower extremity performance tests like the single-leg squat test are per-
formed reliably from week to week, participants were not re-classified
at the second visit (Levinger et al., 2007; Whatman et al., 2011). At
that visit, hip extension and abduction strength were measured with
the MicroFET2™ dynamometer during isometric tests in which partici-
pants exerted maximum force against externally applied resistance.
Hip extension strength was tested with participants in prone and the
knee flexed to approximately 90° and abduction strength was tested
with participants in side-lying, which are standardized positions for
assessing extension and abduction strength, respectively (Hislop and
Montgomery, 2007). The dynamometer was stabilized at the distal
thigh by a strapping belt and an examiner's hand. Two repetitions
were performed and the maximum force was recorded and expressed
as a percentage of the participant's body weight (% BW). Each contrac-
tion was maintained for 5 s and approximately 30 s of rest between
repetitions was provided. EMG data were acquired simultaneously to
establish the amplitude of maximum voluntary isometric contractions
(MVIC) to which subsequent data were normalized. Prior to collecting
EMGs, the participant's skin was cleansed with alcohol. Electrodes
were affixed at standardized locations in parallel with the muscles'
respective lines of action (Criswell, 2011). One electrode pair was
placed over the gluteus maximus at one-half the distance between the
sacrum and greater trochanter. The second electrode pair was placed
over the gluteus medius at one-third the distance between the iliac
crest and greater trochanter. The reference electrode was placed on
the tibial crest. Sixteen retroreflective markers were then placed on
anatomic landmarks to generate the model from which kinematic
data were obtained (Fig. 1). Additional measurements taken to facilitate
kinematic measurements with Vicon's Plug-in-Gait model included
participant height, leg lengths (ASIS to medial malleolus), inter-ASIS
distance, inter-epicondylar distance at the distal thighs and inter-
malleolar distance at the distal shanks. Participants then performed
the 5-repetition single-leg squat test while kinematic and EMG data
were acquired.

Volunteers for study (n = 64)

Excluded (n = 7)
• Reported hip or knee pain in previous 6
months during at least two of the
following activities: running, jumping,
cutting, ascending or descending stairs,
squatting or prolonged sitting (n=4)
• Reported history of ligament injury,
patella dislocation, knee surgery or
other significant trauma to lower
extremity (n = 3)

Met inclusion criteria, screened for single-

leg squat performance (n = 57)

Classified as “Good” Classified as “Fair” Classified as “Poor”

performer (n = 21)
Completed testing
(n = 21)
performer (n = 15)
performer (n = 21)
Excluded (n = 1)
• Missing marker
data during
kinematic analysis
Completed testing
(n = 20)

Fig. 2. Flow chart of recruitment.

Please cite this article as: Hollman, J.H., et al., Frontal and transverse plane hip kinematics and gluteus maximus recruitment correlate with frontal
plane knee kinematics during single-leg squat test..., Clin. Biomech. (2014), http://dx.doi.org/10.1016/j.clinbiomech.2013.12.017
4 J.H. Hollman et al. / Clinical Biomechanics xxx (2014) xxx–xxx

Table 2
Descriptive data in the “good” and “poor” performance groups.

“Good” group “Poor” group 95% CI of the difference P

n = 21 n = 20

Demographic/anthropometric
Age (years) 23.8 (1.8) 24.4 (2.9) −1.0 to 2.0 0.473
Height (cm) 168.2 (7.1) 167.1 (7.1) −5.6 to 3.4 0.624
Mass (kg) 61.3 (8.2) 61.3 (9.6) −5.7 to 5.6 0.989
Body mass index (kg · m−2) 21.6 (2.0) 21.9 (2.3) −1.1 to 1.6 0.690
Physical activity (MET · min · wk−1)
Low activity (b600) 0 (0%) 0 (0%)
Moderate activity (600–3000) 15 (71%) 14 (67%)
High activity (N3000) 6 (29%) 6 (33%) 0.920
Hip strength (% body weight)
Abduction 52.6 (11.9) 54.0 (11.3) −5.9 to 8.7 0.389
Extension 40.8 (10.8) 43.6 (9.9) −3.7 to 9.4 0.706
Hip kinematics (°)
Flexion 42.6 (12.6) 50.1 (8.5) 0.8 to 14.4 0.030⁎
Adduction 13.4 (4.6) 19.7 (5.3) 3.2 to 9.4 b0.001⁎
Medial rotation 2.0 (14.1) 9.0 (9.2) −0.6 to 14.6 0.068
Knee kinematics (°)
Flexion 64.0 (8.3) 64.0 (10.3) −6.0 to 9.0 0.985
Varus 4.2 (12.2) 8.6 (7.2) −2.1 to 10.7 0.181
Medial rotation 6.6 (12.9) 7.7 (12.0) −6.8 to 9.0 0.774
EMG magnitude (% MVIC)
Gluteus maximus 23.8 (11.7) 20.9 (10.7) −10.0 to 4.2 0.413
Gluteus medius 28.3 (17.8) 26.0 (16.1) −13.0 to 8.4 0.666

Continuous variables are presented as mean (SD); ordinal data are presented as discrete numbers in each category (%).
⁎ Statistically significant difference between groups (P b 0.05).

2.4. Data processing
Marker trajectories were filtered with a Woltring quintic spline filter
(20-mm mean square error). Hip and knee joint angles were calculated
with Cardan angles using Vicon's Nexus software, whereby rotations
about orthogonal local axes derived from a neutral standing trial
corresponded to flexion/extension in the sagittal plane, adduction/
abduction (varus/valgus) in the frontal plane and medial/lateral rotation
in the transverse plane, respectively.
Raw EMGs were band-pass filtered between 20 and 450 Hz with a
4th order Butterworth filter and analyzed with Delsys EMGworks
3.7.2.0 software. We processed EMGs with a root mean square algorithm
over 250-ms time constants with sliding windows and normalized data
from the gluteus maximus and gluteus medius to their respective MVIC
trials.
We analyzed 3-dimensional hip and knee angles at the completion
of the eccentric phase of each squat, defined as the point at
which maximum knee flexion was reached. For EMG data, we examined
mean amplitudes over 500-ms epochs preceding maximum knee flexion.
We then computed mean joint angles and mean EMG amplitudes for each
individual across the five repetitions of the single-leg squat test.
2.5. Statistical analysis
Descriptive data were calculated for anthropometric and demo-
graphic characteristics (height, weight, body mass index, and age), hip
extension and abduction strength, hip and knee joint angles and gluteus
maximus and gluteus medius recruitment during the test. Differences
between “good” and “poor” performers were examined with indepen-
dent t-tests for parametric data and χ2 tests for ordinal data.
Second, we used multiple regression to examine relationships
among frontal plane knee angles and 3-dimensional hip angles; gluteus
maximus and gluteus medius recruitment; and isometric hip extension
and abduction strength. We analyzed R2 values and partial correlation
coefficients to examine the relationship among each predictor variable
with frontal plane knee motion, controlling for other variables in the
analysis.

Table 3
Summary of hierarchical regression analysis on the frontal plane knee angle.

Variable R2 ΔR2 B (95% CI) β r partial r P

Constant −11.86
Hip angles (°) 0.865 0.865 b0.01
Frontal plane −0.29 (−0.52 to −0.07) −0.17 0.03 −0.42 0.01
Sagittal plane 0.06 (−0.05 to 0.17) 0.07 −0.10 0.18 0.29
Transverse plane 0.83 (0.72 to 0.93) 0.99 0.92 0.94 b0.01
Hip strength (% BW) 0.873 0.007 0.39
Extension 0.07 (−0.07 to 0.22) 0.07 −0.01 0.17 0.32
Abduction 0.14 (−0.03 to 0.31) 0.16 −0.17 0.28 0.10
Muscle recruitment (% MVIC) 0.895 0.022 0.04
Gluteus maximus 0.13 (0.01 to 0.26) 0.15 −0.06 0.35 0.04
Gluteus medius 0.08 (−0.01 to 0.17) 0.13 0.21 0.29 0.09

Notes: R2 = cumulative proportion of variance in the frontal plane knee angle accounted for by variance in the independent variable; ΔR2 = change in R2; B = unstandardized regression
coefficient (with 95% confidence interval); β = standardized regression coefficient; r = zero-order Pearson product-moment correlation coefficient between the independent variable
and frontal plane knee angle; partial r = partial correlation between the independent variable and frontal plane knee angle, controlling for other independent variables; P = significance
of ΔR2 or of B or β or the partial r.
Using unstandardized regression coefficients, the regression equation is represented as:
Frontal plane knee angle (°) = −11.86°–0.29 (frontal plane hip angle [°]) + 0.83 (transverse plane hip angle [°]) + 0.13 (gluteus maximus recruitment [% MVIC]).

Please cite this article as: Hollman, J.H., et al., Frontal and transverse plane hip kinematics and gluteus maximus recruitment correlate with frontal
plane knee kinematics during single-leg squat test..., Clin. Biomech. (2014), http://dx.doi.org/10.1016/j.clinbiomech.2013.12.017
 J.H. Hollman et al. / Clinical Biomechanics xxx (2014) xxx–xxx 5

We analyzed data with SPSS 21.0 software (IBM Corporation,
Armonk, New York, USA). The threshold for statistical significance for
all statistical tests was α = 0.05. Diagnostic capabilities were used to
assess whether assumptions of linearity, normality and homoscedastic-
ity in the regression analysis were violated. Frontal plane knee data
were linearly associated with the variables included in the regression.
Kolmogorov–Smirnov tests indicated that the variables' distributions
did not depart significantly from normal distributions. Plots of stan-
dardized residuals and standardized predicted values indicated the
assumptions of homoscedasticity were not violated. Variance inflation
factors were less than 3.0, suggesting that multicollinearity was not
problematic.
3. Results
Fifty-seven individuals were screened before 21 were classified as
“good” and 21 as “poor” performers on the test (Fig. 2). Secondary to
technical issues, kinematic data from one participant classified as
“poor” were not acquired so the final sample included 21 “good” and
2011). Methodological differences may potentially account for our con-
trasting results. We examined the magnitude of EMG recruitment in the
gluteus maximus and gluteus medius, for example, whereas Crossley
et al. (2011) only examined gluteus medius onset timing. Furthermore,
differences in our use of the rating scale during the single-leg squat test
may have influenced how participants were categorized. Crossley et al.
(2011) used a consensus panel to categorize participants and did not
assign points for specific components of the test but instead evaluated
the performance as a whole. Their definition of a poor rating reflected
a situation in which intervention would be necessary to address the
hip muscle dysfunction. We believed using the number of movements
that departed from a good standard of performance would be more
objective, but it is possible our threshold for defining poor performance
differed from the Crossley's threshold.
We also examined relationships between frontal plane knee kine-
matics and hip muscle strength and neuromuscular control. Findings
supported our hypothesis that variance in gluteus maximus function
would be associated with variance in knee valgus during the test.

Knee Valgus (degrees) Knee Varus

30
20 “poor” performers. Participant height, weight, BMI, age and physical A
activity levels did not differ between “good” and “poor” performers 20
(Table 2).
Hip kinematics during the test differed between groups. “Poor” 10
performers completed the task with more hip flexion (mean
difference = 7.6°, 95% CI = 0.8° to 14.4°, P = 0.03) and adduction 0
(mean difference = 6.3°, 95% CI = 3.2° to 9.4°, P b 0.01) and tended
to have more medial hip rotation than “good” performers, though the -10
difference was not statistically significant (mean difference = 7.0°,
95% CI = − 0.6° to 14.6°, P = 0.07). In contrast, knee kinematics, hip -20 partial r = 0.939
extension and abduction strength and gluteus maximus and gluteus p < 0.001
medius recruitment did not differ between groups (Table 2). -30
-25 -20 -15 -10 -5 0 5 10 15 20 25
Variance in hip kinematics, hip strength and hip muscle recruitment
accounted for nearly 90% of the variance in frontal plane knee kinemat- Medial Hip Rotation (degrees) Lateral Hip Rotation
ics (Table 3; R2 = 0.895, F7,33 = 40.20, P b 0.01). Three variables con-
Knee Valgus (degrees) Knee Varus

30
tributed significantly to that variance: transverse (partial r = 0.94, B
95% CI = 0.89 to 0.97, P b 0.01) and frontal plane hip motion (partial 20
r = − 0.42, 95% CI = − 0.13 to − 0.64, p = 0.03) and gluteus
maximus recruitment (partial r = 0.35, 95% CI = 0.05 to 0.59, 10
P = 0.04). Holding other variables constant, increased knee valgus
correlated with increased medial hip rotation and adduction and with 0
decreased gluteus maximus recruitment (Figs. 3 & 4).
-10
4. Discussion
-20 partial r = -0.417
We compared hip and knee kinematics as well as gluteus maximus p = 0.013
and gluteus medius function between “good” and “poor” performers -30
on a single-leg squat test. Findings partially supported our hypothesis 5 10 15 20 25
that kinematic performance would differ between groups. “Poor” Hip Adduction (degrees)
performers completed the task with more hip adduction than “good”
Knee Valgus (degrees) Knee Varus

30
performers, which supports the concept that frontal plane hip and C
pelvis control are key elements of performing a single-leg squat. 20
In contrast to our hypothesis, knee angles at the termination of the
squat did not differ between groups, which may imply the performance 10
indicators of the single-leg squat test do not adequately cue the rater to
assess knee motions. Alternatively, that finding may simply reflect the 0
comprehensive nature of the rating scale. Assessing performance during
the single-leg squat test involves the assessment of the trunk and pelvis -10
posture in addition to a general assessment of balance and smoothness
of movement. Participants who were rated poorly on those components -20 partial r = 0.345
of the test did not necessarily perform the test with increased p = 0.043
magnitudes of knee valgus. Moreover, no differences in hip muscle per- -30
10 15 20 25 30 35
formance occurred between “good” and “poor” performers. Neither
extension nor abduction strength differed between groups, nor did glu- Gluteus Maximus Recruitment (% MVIC)
teus maximus or gluteus medius recruitment. These findings contrast Fig. 3. Partial correlations between frontal plane knee kinematics and transverse plane hip
with findings that hip abduction strength was diminished and gluteus kinematics (A), frontal plane hip kinematics (B) and gluteus maximus recruitment
medius recruitment was delayed in “poor” performers (Crossley et al., (C) during the single-leg squat test.

Please cite this article as: Hollman, J.H., et al., Frontal and transverse plane hip kinematics and gluteus maximus recruitment correlate with frontal
plane knee kinematics during single-leg squat test..., Clin. Biomech. (2014), http://dx.doi.org/10.1016/j.clinbiomech.2013.12.017
6 J.H. Hollman et al. / Clinical Biomechanics xxx (2014) xxx–xxx

While transverse and frontal plane hip motions correlated most strong-
ly with frontal plane knee motion, increased knee valgus also correlated
with reduced gluteus maximus recruitment (Fig. 3C). The regression
equation (Table 3) indicates that when other variables are held
constant, a 1.2° increase in medial hip rotation, a 3.4° increase in hip
adduction or a 7.7% MVIC reduction in gluteus maximus recruitment is
associated with a 1° increase in knee valgus during the single-leg
squat test. Assuming the magnitude of muscle recruitment represents
an element of neuromuscular control (Pullman et al., 2000), an implica-
tion is that gluteus maximus neuromuscular control may modulate
frontal plane knee motion during single-leg squats, though the findings
may not generalize to all single-leg tasks.
Two theoretical foundations support our findings. First, the gluteus
maximus and gluteus medius may have distinct relationships with frontal
plane knee kinematics. While Claiborne et al. (2006) reported an associa-
tion between hip abduction strength and knee valgus and asserted that
abduction strength controls frontal plane knee motion, our findings con-
tradict that assertion. We believe there are plausible kinesiologic explana-
tions for our findings. One may suggest that greater abduction strength
can restrict hip adduction and knee valgus motions because the muscle
stabilizes frontal plane hip motion (Khayambashi et al., 2012; Nakagawa
et al., 2008). The pathomechanics about which clinicians are concerned,
however, represents multi-planar phenomena in which excessive move-
ments occur not only in the frontal but also in the transverse plane of mo-
tion. The gluteus medius is primarily a hip abductor but secondarily
assists medial rotation. Its medial rotation moment arm increases as the
hip flexes, particularly beyond 30° (Delp et al., 1999). Per our findings,
the hip flexes at 40–50° during the single-leg squat. An excessively strong
or overly-recruited gluteus medius may act not to prevent hip adduction
but rather to exacerbate medial hip rotation that is coupled to adduction.
Even though trials of hip abduction strengthening for patellofemoral pain
syndrome have shown generally positive outcomes (Khayambashi et al.,
2012; Nakagawa et al., 2008), we hypothesize that promoting gluteus
maximus function may be more relevant to limiting hip adduction, medial
rotation and knee valgus during weightbearing tasks than promoting
abduction strengthening.
The second theoretical foundation is that muscle strength and
neuromuscular control are distinct elements of physiologic function
that influence motor performance. The motor task in our study was
the single-leg squat. Performance was represented by the lower
extremity kinematics. We sought to understand the extent to which
hip extensor and abductor strength and neuromuscular control are
associated with knee kinematics. The question is relevant because,
while many investigators have reported hip strength deficits among
individuals with patellofemoral pain syndrome (Bolgla et al., 2008;
Finnoff et al., 2011; Fulkerson, 2002; Fulkerson and Arendt, 2000;
Ireland et al., 2003), it is evident that hip strengthening in isolation is in-
sufficient to induce changes in frontal plane knee motion (Herman et al.,
2008; Mizner et al., 2008; Snyder et al., 2009). Mizner et al. (2008),
for example, reported that instruction on jump-landing techniques
produced changes in jump-landing performance among collegiate
female athletes, but that hip abduction and extension strength were
not associated with those improvements. Further, while hip strengthen-
ing produces stronger hip muscles, they affect frontal plane hip and
knee kinematics minimally during running or jump-landing activities
(Herman et al., 2008; Snyder et al., 2009). Feedback on motor perfor-
mance induces greater changes in frontal plane kinematics than hip
strengthening alone (Herman et al., 2009). Our finding that gluteus
maximus recruitment during single-leg squats is associated with frontal
plane knee motion supports conclusions that neuromuscular control
rather than strength provides insight into how individuals perform
dynamic lower extremity weightbearing motor tasks (Hollman et al.,

80 80
70
A 70
D
Knee Flexion (°)

60 60
50 50
40 40
30 30
20 20
10 10
0 0
0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14

15 15
B E
10 10
Knee Valgus (°)
Knee Varus

5 5
0 0
-5 -5
-10 -10
-15 -15
0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14

40 40
C F
Gluteus Maximus
EMG (% MVIC)

30 30

20 20

10 10

0 0
0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14
Time (s) Time (s)

Fig. 4. Representative data from a participant who performed the single-leg squat test with neutral to varus frontal plane knee motion (A–C) and a participant with valgus frontal plane knee
motion (D–F). Note the contrast in gluteus maximus recruitment between the two participants. The vertical dashed line segments indicate the times at which peak knee flexion occurred.

Please cite this article as: Hollman, J.H., et al., Frontal and transverse plane hip kinematics and gluteus maximus recruitment correlate with frontal
plane knee kinematics during single-leg squat test..., Clin. Biomech. (2014), http://dx.doi.org/10.1016/j.clinbiomech.2013.12.017
 J.H. Hollman et al. / Clinical Biomechanics xxx (2014) xxx–xxx
2009, 2012; Zazulak et al., 2005). In healthy individuals, one's recruit-
ment of a muscle during a motor task may be more relevant to physio-
logic function and task performance than is absolute muscle strength.
We acknowledge several limitations in this study. Our inter-rater
agreement for classifying performance on the single-leg squat test was
moderately strong (kappa = 0.55) though lower than comparable co-
efficients reported by developers of the rating system (Crossley et al.,
2011). If participants were misclassified as “good” or “poor” performers
on the test or if they performed differently during the testing session,
that may potentially account for the lack of group differences we'd
hypothesized may occur. Second, it's possible that our measurements
were not most representative of hip muscle function. Muscle strength,
for example, is greatest during eccentric contractions. We analyzed
eccentric phases of the single-leg squat. Eccentric strength tests may
have provided greater insight about relationships between knee kine-
matics and hip extension and abduction strength. Similarly, other mea-
sures of neuromuscular control such as EMG onset, peak recruitment or
timing of peak recruitment may have provided different insights than
examining mean recruitment through 500-ms epochs leading to termi-
nation of the squat. Third, participants were healthy, asymptomatic
women. Women with patellofemoral pain syndrome exhibit lower iso-
metric strength than asymptomatic controls (Bolgla et al., 2008; Ireland
et al., 2003). The extent to which gluteus maximus recruitment is asso-
ciated with frontal plane knee motion may differ in participants with
impaired strength. Fourth, surface EMG constrained us from examining
relationships between deep hip external rotators and knee kinematics.
Since minimizing medial hip rotation may be an important component
of limiting knee valgus during single-leg squats, it may be important to
understand how recruitment of the deep hip external rotators is associ-
ated with frontal plane knee motion. In-dwelling EMG could be used to
address that question. Last, it is important to recognize that our study
design does not permit causal inferences. Similarly, while the regression
analysis permits one to examine associations between variables, it does
not imply a cause-and-effect relationship. While reduced gluteus
maximus recruitment as well as increased medial hip rotation and hip
adduction may be associated with increased knee valgus, one should
not interpret that finding to imply that reduced gluteus maximus
recruitment causes knee valgus to increase during a single-leg squat.
Despite these limitations, our finding that increased medial hip
rotation and hip adduction and reduced gluteus maximus recruitment
are correlated with increased knee valgus during a single-leg squat
test is clinically relevant. The finding provides a plausible linkage
between gluteus maximus function and frontal plane knee kinematics.
Future studies should examine how gluteus maximus recruitment can
be enhanced and whether that influences lower extremity kinematics
during dynamic weightbearing tasks.
5. Conclusion
The single-leg squat test used in this study was developed to identify
individuals who may have impaired hip muscle function. In our study, hip
flexion and adduction were increased in female participants identified
as “poor” performers during the test compared to “good” performers.
However, no differences in knee kinematics, hip extension or abduction
strength or in gluteus maximus or gluteus medius recruitment were iden-
tified between the groups. Nevertheless, controlling for hip kinematics,
gluteus maximus recruitment was correlated with frontal plane knee mo-
tion. This finding implies that gluteus maximus neuromuscular control
may modulate frontal plane knee kinematics during single-leg squats.
References
Bolgla, L.A., Malone, T.R., Umberger, B.R., Uhl, T.L., 2008. Hip strength and hip and knee
kinematics during stair descent in females with and without patellofemoral pain
syndrome. J. Orthop. Sports Phys. Ther. 38, 12–18.
Please cite this article as: Hollman, J.H., et al., Frontal and transverse plane hip kinematics and gluteus maximus recruitment correlate with frontal
plane knee kinematics during single-leg squat test..., Clin. Biomech. (2014), http://dx.doi.org/10.1016/j.clinbiomech.2013.12.017
7
Boling, M., Padua, D., Marshall, S., Guskiewicz, K., Pyne, S., Beutler, A., 2010. Gender
differences in the incidence and prevalence of patellofemoral pain syndrome.
Scand. J. Med. Sci. Sports 20, 725–730.
Brindle, T.J., Mattacola, C., McCrory, J., 2003. Electromyographic changes in the gluteus
medius during stair ascent and descent in subjects with anterior knee pain. Knee
Surg. Sports Traumatol. Arthrosc. 11, 244–251.
Claiborne, T.L., Armstrong, C.W., Gandhi, V., Pincivero, D.M., 2006. Relationship between
hip and knee strength and knee valgus during a single leg squat. J. Appl. Biomech.
22, 41–50.
Cowan, S.M., Crossley, K.M., Bennell, K.L., 2009. Altered hip and trunk muscle function in
individuals with patellofemoral pain. Br. J. Sports Med. 43, 584–588.
Craig, C.L., Marshall, A.L., Sjostrom, M., Bauman, A.E., Booth, M.L., Ainsworth, B.E., Pratt, M.,
Ekelund, U., Yngve, A., Sallis, J.F., Oja, P., 2003. International physical activity
questionnaire: 12-country reliability and validity. Med. Sci. Sports Exerc. 35, 1381–1395.
Criswell, E., 2011. Cram's Introduction to Surface Electromyography. Jones and Bartlett
Publishers, Sudbury, MA.
Crossley, K.M., Zhang, W.J., Schache, A.G., Bryant, A., Cowan, S.M., 2011. Performance on
the single-leg squat task indicates hip abductor muscle function. Am. J. Sports Med.
39, 866–873.
Delp, S.L., Hess, W.E., Hungerford, D.S., Jones, L.C., 1999. Variation of rotation moment
arms with hip flexion. J. Biomech. 32, 493–501.
Dixit, S., DiFiori, J.P., Burton, M., Mines, B., 2007. Management of patellofemoral pain
syndrome. Am. Fam. Physician 75, 194–202.
Finnoff, J.T., Hall, M.M., Kyle, K., Krause, D.A., Lai, J., Smith, J., 2011. Hip strength and knee
pain in high school runners: a prospective study. PM R 3, 792–801.
Fredericson, M., Cookingham, C.L., Chaudhari, A.M., Dowdell, B.C., Oestreicher, N.,
Sahrmann, S.A., 2000. Hip abductor weakness in distance runners with iliotibial
band syndrome. Clin. J. Sport Med. 10, 169–175.
Fulkerson, J.P., 2002. Diagnosis and treatment of patients with patellofemoral pain. Am.
J. Sports Med. 30, 447–456.
Fulkerson, J.P., Arendt, E.A., 2000. Anterior knee pain in females. Clin. Orthop. Relat. Res.
372, 69–73.
Herman, D.C., Weinhold, P.S., Guskiewicz, K.M., Garrett, W.E., Yu, B., Padua, D.A., 2008.
The effects of strength training on the lower extremity biomechanics of female
recreational athletes during a stop-jump task. Am. J. Sports Med. 36, 733–740.
Herman, D.C., Onate, J.A., Weinhold, P.S., Guskiewicz, K.M., Garrett, W.E., Yu, B., Padua,
D.A., 2009. The effects of feedback with and without strength training on lower
extremity biomechanics. Am. J. Sports Med. 37, 1301–1308.
Hislop, H.J., Montgomery, J., 2007. Daniels and Worthingham's Muscle Testing:
Techniques of Manual Examination. Saunders Elsevier, St. Louis, MO.
Hollman, J.H., Ginos, B.E., Kozuchowski, J., Vaughn, A.S., Krause, D.A., Youdas, J.W., 2009.
Relationships between knee valgus, hip-muscle strength, and hip-muscle
recruitment during a single-limb step-down. J. Sport Rehabil. 18, 104–117.
Hollman, J.H., Hohl, J.M., Kraft, J.L., Strauss, J.D., Traver, K.J., 2012. Effects of hip extensor
fatigue on lower extremity kinematics during a jump-landing task in women: a
controlled laboratory study. Clin. Biomech. 27, 903–909.
Ireland, M.L., Willson, J.D., Ballantyne, B.T., Davis, I.M., 2003. Hip strength in females with
and without patellofemoral pain. J. Orthop. Sports Phys. Ther. 33, 671–676.
Khayambashi, K., Mohammadkhani, Z., Ghaznavi, K., Lyle, M.A., Powers, C.M., 2012. The
effects of isolated hip abductor and external rotator muscle strengthening on pain,
health status, and hip strength in females with patellofemoral pain: a randomized
controlled trial. J. Orthop. Sports Phys. Ther. 42, 22–29.
Levinger, P., Gilleard, W., Coleman, C., 2007. Femoral medial deviation angle during a one-
leg squat test in individuals with patellofemoral pain syndrome. Phys. Ther. Sport 8,
163–168.
Mizner, R.L., Kawaguchi, J.K., Chmielewski, T.L., 2008. Muscle strength in the lower
extremity does not predict postinstruction improvements in the landing patterns of
female athletes. J. Orthop. Sports Phys. Ther. 38, 353–361.
Myer, G.D., Ford, K.R., Barber Foss, K.D., Goodman, A., Ceasar, A., Rauh, M.J., Divine, J.G.,
Hewett, T.E., 2010. The incidence and potential pathomechanics of patellofemoral
pain in female athletes. Clin. Biomech. 25, 700–707.
Nakagawa, T.H., Muniz, T.B., Baldon, R.D.M., Dias Maciel, C., De Menezes Reiff, R.B., Serrao,
F.B.V., 2008. The effect of additional strengthening of hip abductor and lateral rotator
muscles in patellofemoral pain syndrome: a randomized controlled pilot study. Clin.
Rehabil. 22, 1051–1060.
Powers, C.M., 2003. The influence of altered lower-extremity kinematics on
patellofemoral joint dysfunction: a theoretical perspective. J. Orthop. Sports Phys.
Ther. 33, 639–646.
Prins, M.R., Van Der Wurff, P., 2009. Females with patellofemoral pain syndrome have
weak hip muscles: a systematic review. Aust. J. Physiother. 55, 9–15.
Pullman, S.L., Goodin, D.S., Marquinez, A.I., Tabbal, S., Rubin, M., 2000. Clinical utility of
surface EMG: report of the therapeutics and technology assessment subcommittee
of the American Academy of Neurology. Neurology 55, 171–177.
Riemann, B.L., Lephart, S.M., 2002. The sensorimotor system, part I: the physiologic basis
of functional joint stability. J. Athl. Train. 37, 71–79.
Snyder, K.R., Earl, J.E., O'Connor, K.M., Ebersole, K.T., 2009. Resistance training is accompa-
nied by increases in hip strength and changes in lower extremity biomechanics
during running. Clin. Biomech. 24, 26–34.
Whatman, C., Hing, W., Hume, P., 2011. Kinematics during lower extremity
functional screening tests—Are they reliable and related to jogging? Phys. Ther.
Sport 12, 22–29.
Zazulak, B.T., Ponce, P.L., Straub, S.J., Medvecky, M.J., Avedisian, L., Hewett, T.E., 2005.
Gender comparison of hip muscle activity during single-leg landing. J. Orthop. Sports
Phys. Ther. 35, 292–299.