Available online at www.sciencedirect.

com

Clinical Biomechanics 23 (2008) 779–786

www.elsevier.com/locate/clinbiomech

Effect of foot rotation on knee kinetics and hamstring activation in

older adults with and without signs of knee osteoarthritis
Scott K. Lynn, Patrick A. Costigan *
School of Kinesiology and Health Studies, Queen’s University, PEC, 69 Union St, Kingston, Ontario, Canada K7L 3N6

Received 25 October 2007; accepted 21 January 2008

Abstract

Background. To determine the effects of changing the natural foot progression angle during gait (internal and external foot rotation)
on the knee’s adduction moment, lateral–medial shear force, and the ratio of medial–lateral hamstring muscle activation in those with
signs of knee osteoarthritis and a matched healthy control group.
Methods. Twelve subjects with signs of knee osteoarthritis and 12 matched healthy control subjects were evaluated. A 3D gait analysis
system calculated forces and moments at the knee while the subjects walked in three conditions: (1) normal foot position, (2) external foot
rotation, (3) internal foot rotation. Medial and lateral hamstring EMG data was also collected simultaneously and used to calculate the
medial–lateral hamstring activation ratio during the stance phase of the gait cycle. Repeated measures ANOVAs were used to compare
foot rotation conditions within each group; while between group comparisons were performed in the normal rotation condition only
using t-tests.
Findings. Those with knee osteoarthritis (OA) had an increased late stance knee adduction moment and a decreased medial–lateral
hamstring activation ratio as compared to the healthy control group. Also, external foot rotation decreased the late stance knee adduc-
tion moment, lateral–medial shear force, and hamstring activation ratio. However, internal foot rotation did not increase these measures.
Interpretation. Changes in foot position during gait have the ability to alter both the external loading of the knee joint and hamstring
muscle activation patterns during gait. This may have implication in helping to unload the knee’s articular cartilage.
Ó 2008 Published by Elsevier Ltd.

Keywords: Biomechanics; Locomotion; Adduction moment; Lateral–medial shear force; Medial–lateral hamstring muscle activation

1. Introduction external knee adduction moment. A second knee gait mea-

The external knee adduction moment, which can be esti-
mated at the knee while walking, is a factor in the develop-
ment and progression of knee osteoarthritis (OA) (Lynn
et al., 2007; Miyazaki et al., 2002). A large knee adduction
moment indicates an increased load on the knee’s medial
compartment, while a reduced knee adduction moment
indicates an increased load on the lateral compartment
(Lynn et al., 2007; Weidow et al., 2006). Hence, the gait
profile associated with medial compartment OA, the most
common type of knee OA (Felson, 1998), includes a large
*
Corresponding author. Tel.: +1 613 533 6603; fax: +1 613 533 2009.
E-mail address: pat.costiga@queensu.ca (P.A. Costigan).
sure that has been recently implicated in the development
and progression of knee OA is the lateral–medial shear
force (Astephen and Deluzio, 2005; Lynn et al., 2007). This
is the joint reaction force calculated using inverse dynamics
that acts along the joint surface to push the femur medially
across the tibial plateau. In vitro studies suggest that shear
stress is detrimental to cartilage health (Radin et al., 1991;
Tomatsu et al., 1992) and those with medial compartment
knee OA have been shown to exhibit large lateral–medial
shear forces (Lynn et al., 2007). Clearly, strategies aimed
at reducing the knee adduction moment and lateral–medial
shear forces are warranted.
A simple strategy used by those with medial OA to
reduce the associated pain is to walk with an externally

0268-0033/$ - see front matter Ó 2008 Published by Elsevier Ltd.

doi:10.1016/j.clinbiomech.2008.01.012
780 S.K. Lynn, P.A. Costigan / Clinical Biomechanics 23 (2008) 779–786
rotated foot (Wang et al., 1990). Walking with the foot
rotated reduces the pain and also decreases the measured
knee adduction moment (Guo et al., 2007; Lynn et al.,
2008; Wada et al., 1998) and lateral–medial shear force
(Lynn et al., 2008). This foot rotation strategy is believed
to be effective as those who walk with an increased toe-
out angle during gait were less likely to have their medial
compartment knee OA progress, as compared to those
who walked with less of a toe-out angle (Chang et al.,
2007). Another strategy is to increase the activity of the lat-
eral hamstrings and decrease the activity of the medial
hamstrings during gait (Hubley-Kozey et al., 2006). This
change in muscle activity may produce an internal muscu-
lar abduction (valgus) moment at the knee to counteract
the external gait adduction (varus) moment (Hubley-Kozey
et al., 2006). The rapid application of a valgus or varus
load to the knee is known to produce an increase in medial
and lateral hamstring activity, respectively (Buchanan
et al., 1996), which demonstrates the potential role of the
hamstrings in absorbing frontal plane loads. The mecha-
nism by which those with knee OA alter the activation of
their medial/lateral hamstrings during gait is not clear,
but foot rotation may play a role.
It may be that the two compensations strategies dis-
cussed above (the rotation of the foot and alterations in
medial–lateral hamstring muscle activation) are connected.
Along with knee flexion, the medial and lateral hamstrings
can also produce internal and external rotations of the tibia
(Kendal et al., 2005). Tibial rotation has been shown to
affect the activation of the medial and lateral hamstrings
during maximum isometric knee flexion (Mohamed et al.,
2003); however, to our knowledge, the connection between
foot rotation and hamstrings activation has not been inves-
tigated during gait. Understanding this connection might
help us understand why foot rotation is a common com-
pensation for those suffering from knee OA.
Therefore, this study will examine the changes in ham-
string muscle activation and knee loads during gait with
changes in foot rotation in a knee OA and a healthy control
group. It is hypothesized that as the foot rotates externally,
the knee gait kinetics and hamstring activation ratios will
indicate an unloading of the knee’s medial compartment.
2. Methods
2.1. Participants
Twelve participants with knee OA (six males) were age
and sex matched to a healthy control group with no previ-
ous or current activity restricting condition of the lower
extremities or low back. The average age of the OA sub-
jects was 67.4 years (SD 10.0), while the control subjects
had an average age of 68.7 years (SD 8.4). The OA subjects
had an average height of 1.73 m (SD 0.11), mass of 81.3 kg
(11.8), and BMI of 27.3 (SD 3.8); while the control subjects
had an average height of 1.71 m (SD 0.08), mass of 67.8 kg
(SD 8.6), and BMI of 23.2 (SD 2.8). The OA and control
groups did not differ in age or height but the OA group
was heavier and had a higher BMI (P < 0.05).
All subjects signed a letter of informed consent
approved by the Queen’s University Health Sciences and
Affiliated Teaching Hospitals Research Ethics Board prior
to their participation in the study. All subjects also com-
pleted the Western Ontario and McMaster Universities
Osteoarthritis Index (WOMACÒ) to assess their OA symp-
toms (Bellamy et al., 1988).
2.2. Knee alignment and radiographic evaluation
Each participant also had an anterior–posterior X-ray
taken of their knee. These were used to quantify the radio-
graphic signs of OA and to measure the frontal plane align-
ment of the test limb. In the OA group, the test limb was
the most arthritic limb; while for the control group, it
was the same limb as was tested in their matched OA sub-
ject. From the X-rays, the radiographic signs of OA were
quantified using a modified version of the Scott OA score
(Scott et al., 1993). This version scores only the most
affected tibiofemoral compartment on a scale from 0 (no
signs of OA) to 13 (extreme OA) (Cooke et al., 1999).
Frontal plane knee alignment was also measured from
the X-rays. Since the X-rays did not include the full limb,
alignment was measured as the angle between the femoral
and tibial anatomic axes at the knee. This angle, the femo-
ral shaft-tibial shaft (FSTS) angle, is calculated by drawing
a line that bisects the distal femur and another that bisects
the proximal tibia. Normative data suggests that an ideal
FSTS angle is 5 degrees (i.e. 5 degrees of valgus) (Nguyen
et al., 1989).
2.3. Gait analysis
The gait analysis system that was used consisted of an
optoelectronic motion tracking system (NDI, Waterloo,
Ontario, Canada), and force plates (AMTI, Newton, Mass,
USA) embedded in the walkway to collect kinematic
(100 Hz) and ground reaction force (2000 Hz) data. These
data were used to calculate the net moments and joint reac-
tion forces at the knee, and the angle of the foot during
gait. Participants were fit with marker arrays containing
infra-red light emitting diodes (LEDs). These marker
arrays were secured to the foot, shank, thigh, pelvis (at
the sacrum) and thoracic spine (just below the cervical
spine) with Velcro bands and tape to prevent them from
changing position on the legs and trunk during testing.
A Delsys Bagnoli-8 (Delsys Inc., Boston, MA, USA)
eight channel EMG system was used to collect all EMG
data (differentially amplified with a gain of 1000, bandpass
20–450 Hz, CMRR > 80 dB, input impedance > 1015). Fol-
lowing standard skin preparation, single differential surface
electrodes (DE 2.1 silver surface, inter-electrode distance of
1.0 cm, Delsys Inc., Boston, MA, USA) were applied in
line with the muscles fibers over the muscle bellies of the
biceps femoris (lateral hamstrings) and semimembranosus/
 S.K. Lynn, P.A. Costigan / Clinical Biomechanics 23 (2008) 779–786
semitendinosus (medial hamstrings) and the motor point
was avoided. A National Instruments 12-bit A/D (National
Instruments, Austin, TX, USA) card was used to convert
data from analog to digital form. Specially designed Lab-
View 6.1 (National Instruments, Austin, TX, USA) soft-
ware was used to collect EMG data at 2000 Hz and to
emit a synch pulse when collection was started and stopped.
Once outfitted with all markers and EMG electrodes,
participants walked along the walkway in three conditions:
(1) their natural foot rotation position (NFP), (2) with the
test leg foot externally rotated (EXT) as much as possible
(to where it felt abnormal but not uncomfortable), (3) with
the test leg foot internally rotated (INT) as much as possi-
ble (to where it felt abnormal but not uncomfortable). The
order of the foot conditions were randomized and five good
walking trials were collected for each of the three foot rota-
tion conditions.
Following all gait trials the participant stood in view of
the cameras and a series of one second reference position
trials were collected. A specially designed probe was used
to locate bone landmarks on the segment so that joint cen-
tres could be approximated during processing. Also, a sub-
ject bias and resting EMG trial was collected with the
subject completely relaxed.
Once gait trials were completed the participant per-
formed three repetitions of isokinetic maximum voluntary
hamstring contractions on a Biodex System 3 isokinetic
dynamometer (Biodex Medical Systems Inc., Shirley, NY,
USA) while medial and lateral hamstring EMG data was
collected. These repetitions were performed at 60°/s as this
speed approximates the angular velocities at the knee dur-
ing gait. The EMG data collected during the maximum vol-
untary contractions were used to normalize gait EMG
activation to percent maximum voluntary contraction
(%MVC) in order to provide a physiological basis for
EMG amplitude comparisons between muscles, as has been
done previously (Hubley-Kozey et al., 2006).
2.4. Data processing
Visual 3D (C-Motion Inc., Rockville, MD, USA) was
used to process the walking trial data and estimate the
net external moments and joint reaction forces at the knee.
The EMG data were full wave rectified and filtered using
a 2nd order, low-pass, double-pass Butterworth filter with
a 3 Hz cut-off frequency (Winter, 1990). The peak activa-
tion value achieved during all MVC trials was then identi-
fied for both the medial and lateral hamstrings. Gait EMG
data then had the subject bias/resting EMG levels sub-
tracted off and was then divided by the peak activation dur-
ing MVC trials to give gait EMG waveforms in %MVC
units.
Once all the data was processed, the synchronization
pulse (that defined the start and the end of EMG data col-
lection) was used to align the gait and EMG data in time
and then normalize the data to 100% of the gait cycle.
For each participant, the data were averaged across the five
781
trials for each condition to give one representative curve for
each subject from which variables were extracted for analy-
sis. While all curves used in the ensemble average had sim-
ilar profiles, peak values extracted from the ensemble
average curve may have been slightly reduced if the peaks
for the individual curves did not occur at the same time
point. For the knee adduction moment and lateral–medial
shear force, the early and late stance peaks were used as out-
come measures if they were evident. In some curves these
peaks were not present. For these curves, the value at the
same point in time as the vertical ground reaction force
peak was used, as has been done previously (Guo et al.,
2007; Lynn et al., 2008). To determine the foot progression
angle, the angle of the long axis of the foot segment in the
global (lab) coordinate system, relative to the walking direc-
tions axis, was computed and averaged across the entire
stance phase. The EMG activation across the entire stance
phase for both hamstring muscles was averaged and then
the medial activation level was divided by the lateral to pro-
duce the medial:lateral (M:L) hamstring activation ratio.
2.5. Statistical analysis
Key outcome measures were the early and late stance
knee adduction moment and lateral–medial shear force val-
ues as well as the M:L hamstring activation ratio. Other
variables that were measured included gait speed, foot
rotation angle, frontal plane knee alignment (FSTS angle),
radiographic OA score, and the WOMAC scores.
2.5.1. OA vs. control
Comparisons between the OA and control groups were
performed using independent samples Student t-tests on
the frontal plane knee alignment (FSTS angle), radio-
graphic OA scores, and WOMAC scores. Additional t-tests
compared the OA and control groups on the gait measures
(gait speed, M:L hamstring activation ratio, early and late
stance knee adduction moment and lateral–medial shear
force) in the normal foot rotation condition only.
2.5.2. Foot rotation
The effect of foot rotation on the M:L hamstring activa-
tion ratio, as well as the early and late stance KAM and
LMF variables were tested using repeated measures ANO-
VAs with appropriate Bonferroni corrections to adjust the
critical significance level. If there was a main effect for foot
rotation, all simple comparisons were also tested to deter-
mine the individual differences between foot rotation
conditions.
Repeated measures ANOVAs were also used to test for
differences in gait speed and foot rotation angles across the
three foot rotation conditions. It is known that gait speed
has an effect on knee kinetics (Kirtley et al., 1985) and mus-
cular activity (Hof et al., 2002) so it was important to deter-
mine if there were any differences in speed across foot
rotation conditions and to ensure that our foot rotation
angles were statistically different.
782 S.K. Lynn, P.A. Costigan / Clinical Biomechanics 23 (2008) 779–786

3. Results The average gait curves (knee adduction moment and

lateral–medial shear force) and EMG profiles (medial ham-
Average radiographic OA scores, frontal plane knee string and lateral hamstring) during the stance phase for
alignment (FSTS angle), and WOMAC scores are shown both the OA and control groups across all three foot rota-
in Table 1 for both the OA and control groups. It should tion conditions are presented in Figs. 1 and 2, respectively.
be noted that the OA group had larger radiographic OA, Outcome measures taken from these curves, along with foot
and WOMAC scores (P < 0.05) than the control group, rotation angles and gait speeds, are presented in Table 2.
but there were no differences between groups for frontal The foot progression angle was different across all foot
plane knee alignment. Also, for comparison, the reported rotation conditions for both groups; but there were also
modified Scott radiographic score of 4.4 for our OA group differences in gait speed between conditions. For the OA
(Table 1) would correspond to Kellgren–Lawrence score of group, the internal and external foot rotation conditions
approximately II or III (mild to moderate). produced slower gait speeds than the normal foot rotation
condition; while, for the control group, the gait speed for
the internal rotation condition was slower than the normal
Table 1 foot position condition.
Radiographic and questionnaire scores for OA (n = 12) and control The late stance peak knee adduction moment revealed a
(n = 12) groups main effect for foot rotation in both the OA and control
OA group Control group groups as all subjects were able to reduce the net moment
Radiographic OA scorea,* 4.4 (2.4) 0.3 (0.9) with external rotation. However, the internal foot rotation
Knee alignmentb 0.6 (4.9) 3.1 (2.6) and normal foot rotation condition were inconsistent
WOMACc,* 29.7 (14.0) 3.9 (6.3) across subjects, and hence not different for the late stance
Numbers presented at mean (SD). knee adduction moment peak in either group. The early
*
Significant difference between OA and control group at P < 0.05. stance peak knee adduction moment also did not display
a
Grading of signs of radiographic OA. No OA score = 0, maximum a main effect for foot rotation in the OA group; but the
score = 13.
b
Frontal plane femoral shaft-tibial shaft (FSTS) angle; negative val-
early stance peak knee adduction moment for the internal
ues = varus, positive values = valgus. Normal = 5 degrees valgus. rotation condition was different from both the straight
c
Knee pain, stiffness, physical function questionnaire. No OA symp- and external rotation conditions in the control group. Knee
toms score = 0, and maximum score = 96. adduction moment comparisons between the OA and con-

Fig. 1. Average gait parameters for the OA (A; n = 12) and control (B; n = 12) groups across the stance phase of the gait cycle. Note: KAM = knee
adduction moment; LMF = lateral–medial shear force.
 S.K. Lynn, P.A. Costigan / Clinical Biomechanics 23 (2008) 779–786 783

Fig. 2. Average hamstring EMG activation curves for the OA (A; n = 12) and control (B; n = 12) groups across the stance phase of the gait cycle. Note:
Med. Ham. EMG = medial hamstring muscle activation; Lat. Ham. EMG = lateral hamstring muscle activation.

Table 2
Gait parameters for the OA (n = 12) and control (n = 12) Groups in all three foot rotation positions
OA Control
EXT NFP INT EXT NFP INT
A B C A B
Foot progression angle (deg) 17.1 7.5 4.4 22.5 11.5 2.5C
(8.0) (5.9) (6.4) (5.0) (4.7) (6.4)
Gait speed (m/s) 0.95A 1.04B 0.96A 1.09A,B 1.11A 1.05B
(0.27) (0.28) (0.29) (0.15) (0.13) (0.14)
KAM early stance (Nm/kg) 0.46A 0.45A 0.43A 0.36A 0.37A 0.32B
(0.13) (0.15) (0.15) (0.13) (0.11) (0.12)
KAM late stance (Nm/kg)a 0.31A 0.40B 0.39B 0.19A 0.27B 0.26B
(0.13) (0.14) (0.14) (0.14) (0.12) (0.09)
LMF early stance (N/kg) 0.77A 0.89B 0.84A,B 1.10A 1.10A 1.06A
(0.48) (0.49) (0.47) (0.22) (0.23) (0.30)
LMF late stance (N/kg) 0.62A 0.76B 0.77B 0.90A 1.00A 0.99A
(0.41) (0.45) (0.47) (0.36) (0.33) (0.38)
M:L hamstring activation ratioa 0.51A 0.66B 0.88A,B 0.94A 1.19A 1.87B
(0.34) (0.42) (0.67) (0.41) (0.75) (1.39)
Numbers presented as mean (SD).
For each variable in each population, similar superscripts (A, B, C) indicate no differences (P > 0.05) in simple comparisons between foot rotation
conditions with appropriate Bonferroni corrections.
EXT = external foot rotation position, NFP = normal foot rotation position, INT = internal foot rotation position.
KAM = knee adduction moment, LMF = lateral–medial shear force, M:L = medial:lateral.
a
Significant difference between the OA and control groups for the normal foot rotation (NFP) position (P < 0.05).

trol group for the normal foot rotation position revealed an
increased (P < 0.05) late stance peak for the OA groups but
the early stance peak was not different between groups
(Table 2).
There were differences due to foot rotation in the late
stance lateral–medial shear force for the OA group that
were similar to those seen in the late stance knee adduction
moment as the magnitude of this curve decreased in the
external rotation condition for all subjects, but the internal
and straight foot rotation conditions were not different
(Table 2). The same decrease in magnitude of the late
stance lateral–medial shear force with external rotation
784 S.K. Lynn, P.A. Costigan / Clinical Biomechanics 23 (2008) 779–786
was not present for the control group as subjects were
inconsistent. There were also no differences between the
OA and control groups in the magnitude of either the early
or late stance lateral–medial shear force for the normal foot
rotation condition.
External foot rotation decreased the M:L hamstring
activation ratio while internal rotation increased it,
although not all comparisons between foot rotation condi-
tions reached significance. In the OA group, only the exter-
nal and normal foot rotation conditions were significantly
different; while for the control group, the internal condition
was significantly different than the external and normal
foot rotation conditions, but the external and normal con-
ditions were not statistically different. Also, the comparison
between the OA and control groups revealed that the M:L
hamstring activation ratio was lower for the OA group;
therefore, the OA group favors the activation of the lateral
hamstrings over the medial.
4. Discussion
The main findings of this study are: (1) those with med-
ial knee OA have a decreased M:L hamstring activation
ratio and increased late stance knee adduction moment
during gait as compared to healthy controls; (2) the relative
activation of the medial and lateral hamstrings during gait
was altered with foot rotation; (3) external foot rotation
decreased the late stance knee adduction moment and lat-
eral–medial shear force in both groups, but internal rota-
tion did not increase these two measures.
These results support the theory that the activation of
the medial and lateral hamstrings plays a role in the atten-
uation of frontal plane loading at the knee during gait; as
the OA group had a smaller M:L hamstring activation
ratio and larger late stance peak knee adduction moment
than the control group during the normal foot rotation
condition (Table 2). This smaller M:L hamstring activation
ratio in the OA group indicates that they have an activa-
tion pattern that favours the activation of the lateral over
the medial hamstrings; although this ratio is not an indica-
tor absolute magnitudes of muscle activity, the EMG mag-
nitudes (Fig. 2) also support this theory (i.e. increased
lateral and decreased medial hamstring muscle activity).
Our results agree with previous work where medial OA
subjects were found to have increased lateral hamstring
activity and decreased medial hamstring activity as com-
pared to a healthy control group (Hubley-Kozey et al.,
2006). Therefore, it appears those with medial OA attempt
to produce an internal valgus moment at the knee in an
attempt to balance the increased external varus moment
created during the stance phase of the gait cycle. This
appears to be most useful during early stance, when ham-
string activity is the greatest (Fig. 2) and the position of
the foot does not have an effect on external knee gait kinet-
ics (Guo et al., 2007; Lynn et al., 2008). The role of the
hamstrings in attenuating frontal plane loads at the knee
makes the ability of the individual hamstring muscles to
exert a moment on the knee joint extremely important.
Differences in the torque producing capacities of the
medial and lateral hamstrings have, to our knowledge,
not been previously examined in the literature. It may very
well be the case that there is an inherent imbalance in the
torque producing capacity of the hamstrings that favours
the medial side. This is supported by data examining two
factors that are known to affect the torque producing
capacity of a muscle, the physiological cross sectional area
(PCSA) and the moment arm distance (MAD). It has been
determined that the MAD (Spoor and Vanleeuwen, 1992)
and PCSA (Woodley and Mercer, 2005) of the lateral ham-
strings are approximately half that of the medial ham-
strings. This suggests that, all other things being equal,
with every hamstring contraction throughout a person’s
life, a net muscular varus moment is produced at the knee
joint. This inherent imbalance could be a reason why med-
ial compartment knee OA is much more common than lat-
eral compartment disease in Western cultures (Felson,
1998), and the increased lateral hamstring activation seen
in those with knee OA could be a compensation strategy
attempting to overcome this imbalance to relieve the stress
on the diseased compartment of the knee. Since it appears
that the hamstrings play a role in the frontal plane loading
of the knee, and the inherent M–L hamstring imbalance
would suggest increased stress on the joint’s soft tissues,
interventions aimed at decreasing this imbalance could help
in balancing the loads better between the knee’s medial and
lateral compartments.
The current data also suggests that foot rotation during
gait could help in decreasing the effects of the proposed M–
L hamstring imbalance on OA progression. In both the OA
and control groups, it appears that foot rotation was able
to modify the relative activation of the medial and lateral
hamstrings, as the general trend was a decreased ratio
(favouring the activation of the lateral hamstrings) with
external rotation and an increased ratio (favouring the acti-
vation of the medial hamstrings) with internal rotation,
although not all comparisons reached statistical signifi-
cance. This lack of significance may be due to several fac-
tors such as the small sample size, the differences in gait
speed noted between foot rotation conditions (Table 2),
and the heterogeneity of both subject populations. The
mechanism by which foot rotation alters the activation of
the medial and lateral hamstrings may have to do with
the hamstrings role in the internal and external rotation
of the tibia (Kendal et al., 2005). External rotation of the
tibia may pre-activate the lateral hamstrings and create
increased tension that may enhance its responsiveness to
external loads during the stance phase, while relaxing the
medial hamstrings; and, internal rotation of the tibia would
trigger the opposite response by pre-activating the medial
and relaxing the lateral hamstrings. Therefore, this simple
intervention may help in both decreasing external loads
on the joint’s soft tissues and decreasing the effects of the
proposed M–L hamstring muscular imbalance.
 S.K. Lynn, P.A. Costigan / Clinical Biomechanics 23 (2008) 779–786
Differences in external knee gait kinetics with foot rota-
tion also provided some interesting results. As expected,
external rotation of the foot decreased the late stance knee
adduction moment in both the OA and Control popula-
tion; and this agrees with the literature (Guo et al., 2007;
Lynn et al., 2008). However, this peak was not different
between the internal and normal foot rotation conditions
for either group. This contradicts the literature as internal
foot rotation has been shown to increase the knee adduc-
tion moment during late stance (Lynn et al., 2008); how-
ever, this was determined in healthy young subjects. It
has been suggested that internal rotation shifts the loads
more onto the knee’s medial compartment (Lynn et al.,
2008), which may create and uncomfortable environment
for older adults and cause them to make further compensa-
tions in their gait patterns. This is supported by the fact
that both the OA and control groups walked at a slower
velocity with their foot internally rotated then they did in
their normal foot rotation condition (Table 2). Also, the
fact that the early stance knee adduction moment peak
was not different between the OA and control groups in
the normal foot rotation condition can be explained by
the fact that the knee alignments were not different between
the groups either (Table 1), as this peak is best predicted by
the static alignment of the joint (Hurwitz et al., 2002).
Foot rotation had a similar effect on the lateral–medial
shear force as it did for the knee adduction moment in the
OA group; as the external rotation decreased the magnitude
of the force in late stance as compared to the normal foot
position condition, but there was no change with internal
rotation. This decrease in the magnitude of the lateral–med-
ial shear force at the knee with external foot rotation may
help in taking the stress off the medial compartment articu-
lar cartilage (Lynn et al., 2007, 2008) and hence, be a mech-
anism that may help to explain why those who walk with the
foot externally rotated have a decreased likelihood of OA
progression (Chang et al., 2007). A similar pattern was pres-
ent in the Control group yet there was no main effect of foot
rotation in this group. The reason foot rotation did not
affect the lateral–medial shear force in the control group
should be examined further as it may be a function of sev-
eral factors that such as muscle strength and balance, gait
speed, static and dynamic knee alignment, and knee laxity
in the frontal plane. It is known that shear forces are detri-
mental to cartilage health (Radin et al., 1991; Tomatsu
et al., 1992) yet this lateral–medial shear force during gait
has not been studied as extensively as the knee adduction
moment to date; therefore, its role in the development
and progression of knee OA is not well understood.
Differences in gait speed have an effect on both the exter-
nal loading of the knee joint (Kirtley et al., 1985) and the
muscle activity of the lower limb (Hof et al., 2002),
although it is not clear how the differences in gait speed
may have affected the results of this current work. Chiu
and Wang (2007) demonstrated that increases in walking
speed are associated with increased hamstring muscle acti-
vation but EMG data was only recorded from the biceps
785
femoris (lateral hamstrings). It is unclear how differences
in gait speed would affect the ratio of medial and lateral
hamstrings activation.
The effect of gait speed on the frontal plane knee gait
kinetics is also unclear. Studies investigating the relation-
ship between gait speed and the peak knee adduction
moment range from finding a weak/poor relationship
(Kirtley et al., 1985) to a positive correlation that changes
depending on the severity of OA in the population tested
(Mundermann et al., 2004); yet these studies did not per-
form separate analyses for the early and late stance peaks.
Those that did perform separate analyses for the early and
late stance peaks have also presented some conflicting
results (Thorp et al., 2006; Oakley, 2000). Future work
should attempt to clarify this relationship and also deter-
mine the combined effects of foot rotation and changes in
gait speed on external frontal plane knee gait kinetics and
hamstring muscle activation. Some of the aberrant finding
in this work may be due to the differences in gait speed
between conditions and therefore, it is essential that this
relationship be clarified.
The results of this current work suggest that the ratio of
medial–lateral hamstring muscle activity may play a role in
the attenuation of frontal plane loading at the knee and
that the position of the foot during the stance phase of
the gait cycle also has the potential to alter both the mus-
cular and external loads applied to the knee joint. This may
have implications in the development of simple, non-inva-
sive interventions aimed at shifting the loads off of over-
loaded or diseased knee compartments.
Acknowledgements
The authors would like to thank Queen’s University for
their financial support and the Motor Performance Lab in
the School of Rehabilitation Therapy at Queen’s for use of
their lab and equipment for data collection. We would also
like to thank Elizabeth Sled for her help in subject recruit-
ment and data collection.
References
Astephen, J.L., Deluzio, K.J., 2005. Changes in frontal plane dynamics
and the loading response phase of the gait cycle are characteristic of
severe knee osteoarthritis application of a multidimensional analysis
technique. Clin. Biomech. 20, 209–217.
Bellamy, N., Buchanan, W.W., Goldsmith, C.H., Campbell, J., et al.,
1988. Validation-study of Womac – a health-status instrument for
measuring clinically important patient relevant outcomes to antirheu-
matic drug-therapy in patients with osteoarthritis of the hip or knee. J.
Rheumatol. 15, 1833–1840.
Buchanan, T.S., Kim, A.W., Lloyd, D.G., 1996. Selective muscle
activation following rapid varus. Med. Sci. Sports Exer. 28, 870–876.
Chang, A., Hurwitz, D., Dunlop, D., Song, J., Cahue, S., Hayes, K., et al.,
2007. The relationship between toe-out angle during gait and
progression of medial tibiofemoral osteoarthritis. Ann. Rheum. Dis.
66 (10), 1271–1275.
Chiu, M.C., Wang, M.J., 2007. The effect of gait speed and gender on
perceived exertion, muscle activity, joint motion of lower extremity,
786 S.K. Lynn, P.A. Costigan / Clinical Biomechanics 23 (2008) 779–786
ground reaction force and heart rate during normal walking. Gait
Posture 25, 385–392.
Cooke, T.D.V., Kelly, B.P., Harrison, L., Mohamed, G., et al., 1999.
Radiographic grading for knee osteoarthritis. A revised scheme that
relates to alignment and deformity. J. Rheumatol. 26, 641–644.
Felson, D.T., 1998. Epidemiology of osteoarthritis. In: Brandt, K.D.,
Doherty, M., Lohmander, L.S. (Eds.), Osteoarthritis, first ed. Oxford
University Press, Oxford, pp. 13–22.
Guo, M., Axe, M.J., Manal, K., 2007. The influence of foot progression
angle on the knee adduction moment during walking and stair
climbing in pain free individuals with knee osteoarthritis. Gait Posture
26, 436–441.
Hof, A.L., Elzinga, H., Grimmius, W., Halbertsma, J.P.K., 2002. Speed
dependence of averaged EMG profiles in walking. Gait Posture 16, 78–
86.
HubleyKozey, C.L., Deluzio, K.J., Landry, S.C., McNutt, J.S., et al.,
2006. Neuromuscular alterations during walking in persons with
moderate knee osteoarthritis. J. Electromyogr. Kines. 16, 365–378.
Hurwitz, D.E., Ryals, A.B., Case, J.P., Block, J.A., et al., 2002. The knee
adduction moment during gait in subjects with knee osteoarthritis is
more closely correlated with static alignment than radiographic disease
severity, toe out angle and pain. J. Orthop. Res. 20, 101–107.
Kendal, F.P., McCreary, E.K., Provance, P.G., Rodgers, M.M., et al.,
2005. Muscles – Testing and Function with Posture and Pain, fifth ed.
Lippincott Williams & Wilkins, Philadelphia.
Kirtley, C., Whittle, M.W., Jefferson, R.J., 1985. Influence of walking
speed on gait parameters. J. Biomed. Eng. 7, 282–288.
Lynn, S.K., Reid, S.M., Costigan, P.A., 2007. The influence of gait pattern
on signs of knee osteoarthritis in older adults over a 5–11 year follow-
up period: a case study analysis. Knee 14, 22–28.
Lynn, S.K., Kajacs, T., Costigan, P.A., 2008. The effect of internal and
external foot rotation on the adduction moment and lateral–medial
shear force at the knee during gait. J. Sci. Med. Sport. doi:10.1016/
j.jsama.2007.03.00.
Miyazaki, T., Wada, M., Kawahara, H., Sato, M., et al., 2002. Dynamic
load at baseline can predict radiographic disease progression in medial
compartment knee osteoarthritis. Ann. Rheum. Dis. 61, 617–622.
Mohamed, O., Perry, J., Hislop, H., 2003. Synergy of medial and lateral
hamstrings at three positions of tibial rotation during maximum
isometric knee flexion. Knee 10, 277–281.
Mundermann, A., Dyrby, C.O., Hurwitz, D.E., Sharma, L., et al., 2004.
Potential strategies to reduce medial compartment loading in patients
with knee osteoarthritis of varying severity – reduced walking speed.
Arthritis Rheum. 50 (4), 1172–1178.
Nguyen, C., Bryant, J.T., Cooke, T.D.V., Chow, D., 1989. Alignment and
geometry of the normal knee in stance. J. Bone Joint Surg. 71-B,
346.
Oakley, P.A., 2000. Changes in Three-dimensional Knee Kinematics and
Kinetics within Normative Gait Speeds. Unpublished Masters Thesis.
Queen’s University, Kingston, ON, Canada.
Radin, E.L., Burr, D.B., Caterson, B., Fyhrie, D., et al., 1991.
Mechanical determinants of osteoarthrosis. Semin. Arthritis.
Rheum. 21, 12–21.
Scott, W.W., Lethbridgecejku, M., Reichle, R., Wigley, F.M., et al., 1993.
Reliability of grading scales for individual radiographic features of
osteoarthritis of the knee – the baltimore-longitudinal-study-of-aging
atlas of knee osteoarthritis. Invest. Radiol. 28, 497–501.
Spoor, C.W., Vanleeuwen, J.L., 1992. Knee muscle moment arms from
MRI and from tendon travel. J. Biomech. 25, 201–206.
Thorp, L.E., Sumner, D.R., Block, J.A., Moisio, K.C., et al., 2006. Knee
joint loading differs in individuals with mild compared with moderate
medial knee osteoarthritis. Arthritis Rheum. 54, 3842–3849.
Tomatsu, T., Imai, N., Takeuchi, N., Takahashi, K., et al., 1992.
Experimentally produced fractures of articular-cartilage and bone –
the effects of shear forces on the pig knee. J. Bone Joint Surg. 74A,
457–462.
Wada, M., Imura, S., Nagatani, K., Baba, H., et al., 1998. Relationship
between gait and clinical results after high tibial osteotomy. Clin.
Orthop. 354, 180–188.
Wang, J.W., Kuo, K.N., Andriacchi, T.P., Galante, J.O., 1990. The
influence of walking mechanics and time on the results of proximal
tibial osteotomy. J. Bone Joint Surg. 72A, 905–909.
Weidow, J., Tranberg, R., Saari, T., Karrholm, J., 2006. Hip and knee
joint rotations differ between patients with medial and lateral knee
osteoarthritis: gait analysis of 30 patients and 15 controls. J. Orthop.
Res. 24, 1890–1899.
Winter, D.A., 1990. Biomechanics and Motor Control of Human
Movement, second ed. Wiley Interscience, New York, p. 205.
Woodley, S.J., Mercer, S.R., 2005. Hamstring muscles: architecture and
innervation. Cells Tissues Organs 179, 125–141.