External loading of the knee joint during

running and cutting maneuvers

THOR F. BESIER, DAVID G. LLOYD, JODIE L. COCHRANE, and TIMOTHY R. ACKLAND
Department of Human Movement & Exercise Science, University of Western Australia, Perth, AUSTRALIA

ABSTRACT
BESIER, T. F., D. G. LLOYD, J. L. COCHRANE, and T. R. ACKLAND. External loading of the knee joint during running and cutting
maneuvers. Med. Sci. Sports Exerc., Vol. 33, No. 7, 2001, pp. 1168 –1175. Purpose: To investigate the external loads applied to the
knee joint during dynamic cutting tasks and assess the potential for ligament loading. Methods: A 50-Hz VICON motion analysis
system was used to determine the lower limb kinematics of 11 healthy male subjects during running, sidestepping, and crossover cut.
A kinematic model was used in conjunction with force place data to calculate the three-dimensional loads at the knee joint during stance
phase. Results: External flexion/extension loads at the knee joint were similar across tasks; however, the varus/valgus and internal/
external rotation moments applied to the knee during sidestepping and crossover cutting were considerably larger than those measured
during normal running (P ⬍ 0.05). Sidestepping tasks elicited combined loads of flexion, valgus, and internal rotation, whereas
crossover cutting tasks elicited combined loads of flexion, varus, and external rotation. Conclusion: Compared with running, the
potential for increased ligament loading during sidestepping and crossover cutting maneuvers is a result of the large increase in
varus/valgus and internal/external rotation moments rather than any change in the external flexion moment. The combined external
moments applied to the knee joint during stance phase of the cutting tasks are believed to place the ACL and collateral ligaments at
risk of injury, particularly at knee flexion angles between 0° and 40°, if appropriate muscle activation strategies are not used to counter
these moments. Key Words: LIGAMENT INJURY, SIDESTEPPING AND CUTTING, KNEE JOINT LOADS

T
he etiology of noncontact knee ligament injury is less
certain than that for contact or impact injuries that are
common to football or skiing (22). In the case of the
anterior cruciate ligament (ACL), noncontact injuries appear
to have a common feature in that they typically involve
sudden changes in direction combined with acceleration or
deceleration of the body (22). Cross and colleagues (5) have
recognized that sidestep cutting maneuvers are a common
mechanism of noncontact or isolated ACL rupture. To un-
derstand the mechanisms behind noncontact ACL injury, it
is imperative to measure the external loads applied to the
knee during tasks that challenge the integrity of the knee
joint. “Only when the causal relations between applied
forces and resultant injury are established and understood
can appropriate programs of intervention and prevention be
designed and implemented” (25). To date, there is little
documentation regarding the external loads applied to the
knee during sport-specific movements that involve rapid
changes in direction. Measuring joint loads at the knee
during tasks such as sidestepping and crossover cutting
therefore appears a logical step in understanding the etiol-
ogy of noncontact knee ligament injury.
Markolf et al. (12,13) and Wascher et al. (24) measured
the loads on ACL and PCL in fresh frozen cadaveric knees
under combined loads applied to the tibia throughout a range
of knee angles. The combination of loads included flexion,
0195-9131/01/3307-1168/$3.00/0
MEDICINE & SCIENCE IN SPORTS & EXERCISE®
Copyright © 2001 by the American College of Sports Medicine
Received for publication May 2000.
Accepted for publication October 2000.
1168
adduction/abduction (varus/valgus, respectively), and inter-
nal/external rotation. A combination of externally applied
internal rotation and anterior tibial force resulted in the
highest load on the ACL. Pure internal rotation moments, as
well as varus-valgus moments in conjunction with anterior
tibial force produced large loads on the ACL. Markolf et al.
(12) also found that the ACL experienced large loads at
extended knee angles, particularly between 0° and 20° of
knee flexion. This finding is in agreement with other studies
that have measured or estimated ACL load at different knee
flexion angles (2,19).
Movements such as sidestepping and crossover cutting
are expected to involve various combinations of external
flexion/extension (FE), varus/valgus (VV), and internal/ex-
ternal rotation (IE) moments at the knee and thus place
different loads on the soft tissues and supporting structures
compared with straight running. The purpose of this inves-
tigation was to quantify the three-dimensional external knee
moments and knee flexion angle during dynamic sidestep-
ping and crossover cutting, and suggest how knee loading
during cutting movements might contribute to increased risk
of injury. These parameters were chosen based on the
known interactions between the external load, knee flexion
angle, and the potential for ligament loading (2,13).
METHODS
Subjects. Eleven healthy, male soccer players without
history of lower limb injury volunteered for this investiga-
tion (mean ⫾ SD age: 21.3 ⫾ 3.4 yr; height: 179.4 ⫾ 7.8
cm; mass: 74.1 ⫾ 7.1 kg). Subjects with previous ankle or
knee injury (sustained within the previous year) were
excluded from the study. Amateur soccer players from two
first grade competition teams were selected for participation
as it was expected that this population would be familiar
with performing sidestepping and crossover cutting tasks.
Before data collection, the testing protocols were explained
to the subjects, and their written informed consent obtained.
Experimental design. Subjects were asked to perform
repeated trials of four tasks in the gait laboratory during
three sessions conducted over 3 wk. Subjects were asked to
step off both right and left legs during the initial familiar-
ization trials to determine which foot they preferred to step
from. After these initial trials, it was found that all subjects
preferred performing the tasks from their right leg and were
asked to perform the tasks in bare feet to negate any effects
due to wearing different running shoes. A thin carpet was
laid on the running surface to prevent reflections being
picked up by the motion analysis cameras and provided
adequate friction between the bare foot and the ground. A
few days before the start of testing, subjects completed a 1-h
training session to familiarize themselves with each of the
tasks and the testing procedures. The tasks were performed
in a random order to account for fatigue, with 10 trials of
each maneuver recorded, giving a total of 40 trials per
session. One-minute intervals were also given between each
trial to reduce the effects of fatigue.
The testing protocol was chosen to ensure consistency
among subjects in regard to the cutting angle and speed
maintained. Previous studies on sidestepping or crossover
cutting have not restricted the cutting angle or speed of task
execution, which could result in a large variation in the
loads experienced at the knee joint (3,4). McLean et al. (15)
limited their sidestepping tasks within a range of 35– 60°
from the direction of travel, whereas Neptune et al. (16)
chose 45° as the desired direction of travel during their
investigation of sidestepping maneuvers. It was decided to
examine the loads at the knee during two sidestepping tasks,
one task being relatively easy to perform (30° from the
direction of travel) and the other task being more difficult to
perform (60° from the direction of travel). The crossover cut
was also included in the protocol, as it was believed that this
task would produce quite different patterns of external load-
ing compared with the sidestepping tasks.
To summarize, the four tasks were a straight run (RUN),
sidestep to 30° (S30) and 60° (S60) stepping left off the right
foot, and a crossover cut to 30° (XOV) stepping right off the
right foot. For each trial, lights on a target board indicated
the task direction and tape was placed on the floor to
indicate the cutting angle required (see Fig. 1). Subjects
were instructed to maintain their running speed as much as
possible throughout the task. Infrared timing gates were
used to monitor the approach running speed (see Fig. 1),
which was delimited to 3 m·s-1 (~10 km·h-1). The speed was
chosen to be as fast as possible within the limits of the
50-Hz motion analysis system. Using a similar method to
that proposed by Antonsson and Mann (1), a Fourier anal-
ysis of the ground reaction forces recorded during each trial
confirmed that the 98% of the frequency content of the
cutting tasks were at 10 Hz or below. This was considered
LOADING AT THE KNEE DURING CUTTING TASKS
FIGURE 1—Target board and gait laboratory set up. The lights on the
target board correspond to the desired direction of travel. Note that
these lines of travel are also marked on the floor with tape. The force
plate lies at the apex of these lines.
to provide an adequate signal-to-noise ratio to reconstruct a
smooth signal from the 50-Hz motion analysis system.
Data collection. Retro-reflective markers were fixed
to lower limb landmarks to record three-dimensional lower
limb movements by using a six-camera, 50-Hz VICON
motion analysis system (Oxford Metrics Ltd., Oxford,
United Kingdom). The VICON Clinical Manager (VCM)
marker set was used (6,9), which consisted of 13 lower limb
markers placed on: left and right anterior superior iliac
spines (ASIS); sacrum; left and right knee joint centers; both
lateral malleoli; the head of the second metatarsal on each
foot; left and right mid-shank, in line with the malleolus and
knee joint markers; and left and right mid-thigh, in line with
the greater trochanter and knee joint center. A static trial
was collected before dynamic testing. This static trial in-
cluded markers on each heel (in line with the metatarsal
marker) to complete the foot segment and a knee alignment
device (KAD) to determine the correct knee joint center and
knee joint axis alignment. Tibial torsion (defined as the
offset between knee and ankle FE axes) was also measured
and used as input into the kinematic model (as required for
VCM software).
Ground reaction forces (GRF) were recorded at 2000 Hz
using a 1200 ⫻ 600 mm force-plate (Advanced Mechanical
Technology Inc., Watertown, MA). Subjects were aware of
the position of the plate to ensure they were performing the
tasks at the same position but were instructed to look
straight ahead when performing each task so that they did
not “target” the force-plate. To this end, a starting marker
was used to alter the subjects’ run-up distances to ensure
that the right foot landed in the center of the plate without
the need to target.
Data analysis. Kinematic data and inverse dynamic
calculations were performed using VCM software (Oxford
Metrics Inc.). This software was used as it employed stan-
dard and accepted methods for processing gait data (6). This
method has also been used previously to perform three-
dimensional analyses of running gait (17,18). Lower limb
segment trajectories and ground reaction force data were
filtered and interpolated simultaneously within VCM by
Medicine & Science in Sports & Exercise姞 1169

FIGURE 2—Schematic of the three stages of stance phase, determined
using resultant ground reaction force. WA, weight acceptance; PPO,
peak push off; FPO, final push off.
fitting a Bezier spline to the data (VCM, Oxford Metrics
Inc.). Joint kinetic data were calculated using an inverse
dynamic analysis as described by Kadaba et al. (9) and
Davis et al. (6). The joint moments referred to in this paper
are the external moments applied to the joint. For example,
an external flexion load will tend to flex the knee.
Knee flexion angle, GRF, and three-dimensional joint ki-
netic data were normalized to stance phase by fitting a cubic
spline to the data using Microsoft Excel® (Microsoft Corp.)
and interpolating the spline to 30 points. Knee moments and
flexion angle were averaged across three phases of stance,
based upon the magnitude of the resultant GRF (vector sum-
mation of Fx, Fy, and Fz). Figure 2 illustrates the three phases
during stance, which were determined as follows:
1. Weight acceptance (WA)—from heel strike to the first
trough in the resultant GRF.
2. Peak push off (PPO)—10% either side of the peak
resultant GRF, and
3. Final push off (FPO)—last 15% of stance.
The initial speed of each task was calculated from the
infrared timing gates and monitored to ensure that each task
was performed at the same speed of ~3 m·s-1. For the
determination of the speed and cutting angle throughout the
performance of the maneuver, the kinematics of the pelvic
center were determined for one stride. One stride was de-
fined from when the thigh of the step leg was vertical before
foot contact to when the same thigh was vertical after
contact with the force plate. The average speed maintained
throughout the maneuver was calculated using the x and y
displacements of the pelvic center (anterior/posterior and
medio/lateral displacements, respectively), i.e.,
Speed ⫽ ((xi ⫺ xi⫺1)2 ⫹ (yi ⫺ yi⫺1)2)1/2/(ti⫺ti⫺1), where i
⫽ ith time point (i.e., ti).
and the cutting angle calculated as follows:
Cutting angle ⫽ tan⫺1[(yi ⫺ yi⫺1)/(xi ⫺ xi⫺1)], where i ⫽ ith time point.
1170 Official Journal of the American College of Sports Medicine
The speed and cutting angle used in the final analysis was
the average of those parameters determined across the final
five time points of the stride.
Reliability and validity of gait data. Coefficients of
multiple determination (R2) were calculated for GRF, knee
flexion angle, and three-dimensional moment data across
the stance phase to determine the reliability within and
between testing sessions (9).
Statistical design. The data assembled for the statis-
tical analysis were the three-dimensional joint moments and
knee flexion angles in each of the stance phases (WA, PPO,
and FPO) and the running speed and cutting angle achieved.
All statistical analyses were performed using Datadesk®
statistical software (Data Description Inc., Ithaca, NY) and
significance indicated with P ⬍ 0.05. Scheffé post hoc tests
were conducted to determine significant interactions and
differences among task.
It was found that week 1 had less repeatability in mea-
sured ground reaction forces and knee joint moments com-
pared with the second and third testing sessions, with lower
R2 values. A two-way ANOVA (stance phase ⫻ week)
indicated that there were differences in the calculated knee
joint moments, knee flexion angle, and cutting angle
achieved during each of the cutting tasks across weeks.
There were significant differences (P ⬍ 0.05) in many of the
measured VV and IE moments, and knee flexion angles
between weeks 1–2 and weeks 1–3. However, there were no
differences in any parameters between week 2 and week 3.
These results were probably due to learning effects and it
was decided to pool data from week 2 and week 3 and treat
data from week 1 as a training session. Subsequently, the
data parameters pooled from week 2 and week 3 were
analyzed using a two-way ANOVA (stance phase ⫻ task).
RESULTS
Repeatability of gait data. The GRF and joint mo-
ments measured during the running and cutting tasks were
very repeatable between and within week 2 and week 3.
There was high repeatability between testing sessions, es-
pecially in the IE moments (Table 1a). The XOV task was
performed with the most variation across testing sessions;
however, the data were still repeatable from week to week.
Repeatability within each testing session was also high
(Table 1b). The speeds and cutting angles measured during
each task were similar across weeks 2 and 3, suggesting that
no learning had occurred between these testing sessions.
Greater variation in the GRF occurred in the frontal
(mediolateral) plane during running, compared with the
sidestepping tasks. This was not unexpected, however,
owing to the small forces in the medial-lateral directions
during the RUN task, which could be affected by small
trial-to-trial variations. This is in comparison to the cutting
tasks that elicited large forces in the medial-lateral direc-
tions and would not be affected as much by the small
variations between trials. Moments applied to the knee in FE
showed the least amount of variation within each testing
session (mean R2 ⫽ 0.96 ⫾ 0.03), with the VV and IE loads
http://www.acsm-msse.org
TABLE 1a. Between testing session mean (SD) coefficients of multiple determination (R2) for preplanned running and cutting tasks.
Direction S60
FE 0.89 (0.04)
Moment VV 0.93 (0.06)
IE 0.99 (0.01)
having greater variation (average VV R2 ⫽ 0.87 ⫾ 0.15 and
average IE R2 ⫽ 0.84 ⫾ 0.09). Repeatability of the knee FE
angle was very high between and within testing sessions (R2
⫽ 0.94 ⫾ 0.04).
Task performance measures: speed and cutting
angle. The run up speed measured between the timing
gates was similar for all subjects across all tasks and was
consistent for different testing sessions (mean ⫽ 3.13 ⫾
0.27 m·s-1). The average speed throughout the maneuver
was dependent on the task being performed. The S30 and
XOV were ~0.2 m·s-1 slower than the RUN (P ⬍ 0.05) and
the S60 task was slower than the RUN by ~0.4 m·s-1 (P ⬍
0.05).
The desired cutting angle for the S30 and XOV were
attained, albeit with a reduction in speed. However, the
angle of 60° for the S60 task was not attained, even with a
significant reduction in running speed (mean cutting angle
⫽ 56.0 ⫾ 4.4°). This confirmed that the S60 task was
difficult to perform.
Flexion/extension moments. During PPO the knee
was subject to large flexion moments, with little difference
between tasks (Fig. 3). This load was equivalent to 150 N·m
for a 75-kg person at PPO. The sidestepping and crossover
cutting tasks elicited greater flexion loads compared to the
RUN; however, this was only significant for the S30 during
PPO (P ⬍ 0.05). The external flexion loads applied to the
knee at FPO were negligible compared to the moments
measured at WA and PPO.
Varus/valgus moments. Significant differences (P ⬍
0.05) in VV moments were found between the cutting tasks
compared to the RUN at all stages of the stance phase (Fig.
4). Both sidestepping tasks placed an external valgus load
on the knee at WA and FPO, whereas a varus load was
applied to the knee during the RUN and XOV tasks at
similar stages of stance. The valgus load during the S60 task
was approximately 2 and 6 times greater than the varus load
experienced during the RUN at WA and FPO, respectively
(~40 N·m for a 75-kg person). The varus load during the
XOV was more than twice that experienced during the RUN
at similar phases of the stance phase (~40 N·m for a 75-kg
person). At PPO, net varus loads were experienced in all
tasks and were generally larger than those at WA and FPO
TABLE 1b. Within session mean (SD) coefficients of multiple determination (R2) for ground reaction forces (GRF) and external knee moments.
Parameter Direction S60
AP 0.96 (0.02)
GRF ML 0.94 (0.03)
Vert 0.95 (0.03)
FE 0.96 (0.02)
Moment VV 0.86 (0.14)
IE 0.83 (0.08)
AP, Anteroposterior GRF; ML, Mediolateral GRF; Vert, Vertical GRF; FE, Flexion/extension moment; VV, Varus/valgus moment; IE, Internal/external rotation moment; S60, Sidestep to
60°; S30, Sidestep to 30°; RUN, Straight run; XOV, Crossover cut to 30°.
LOADING AT THE KNEE DURING CUTTING TASKS
S30 RUN XOV
0.92 (0.04) 0.90 (0.05) 0.84 (0.10)
0.91 (0.09) 0.88 (0.10) 0.81 (0.16)
0.99 (0.01) 0.99 (0.01) 0.98 (0.03)
(Fig. 4). The RUN was again significantly different from the
other tasks (P ⬍ 0.05).
Note the large standard deviations in the VV loads at PPO
(Fig. 4). Upon closer inspection, it was found that six of the
subjects actually experienced a valgus load at PPO during
the S60 and S30 tasks (mean: ⫺0.31 N·m·kg-1). The other
five subjects experienced a large varus load at the knee
(mean: 0.62 N·m·kg-1) such that the net moment for all
subjects was in a varus direction and was smaller than that
of the RUN (P ⬍ 0.001). Differences between these groups
were significant (P ⬍ 0.01), with no intermediate group
apparent. Therefore, these groups shall be referred to as
“valgus group” or “varus group,” accordingly.
There was also a significant difference in the speed at
which the sidestepping tasks were performed between the
valgus and varus groups. The valgus group performed the
S60 task 0.25 m·s-1 slower than the varus group (P ⬍ 0.01).
There were no significant differences in the cutting angles
attained by both groups although there was a trend that
showed the valgus group did not reach the desired 60°
sidestep cutting angle as well as the varus group, with the
result limited by the small numbers in each group. Interest-
ingly, there were no differences between these two groups
for FE and IE moments.
Internal/external rotation moments. Significant
differences (P ⬍ 0.05) in the IE moments measured at the
knee were found between the S60, S30, and XOV tasks
compared with the RUN (Fig. 5). The differences between
the tasks were significant for all stages of the stance phase.
The IE moments experienced during the cutting tasks appear
small when expressed as a percentage of body weight;
however, when compared with the moments measured dur-
ing the RUN, they are quite substantial. During WA, the
sidestepping tasks elicited an internal rotation moment up to
4 times the magnitude of the external rotation moment
experienced during the RUN (P ⬍ 0.001). The external
rotation load applied to the knee during the XOV was more
than twice the load experienced during the RUN (P ⬍
0.001). The internal rotation loads experienced at PPO dur-
ing the sidestepping tasks were up to 5 times the load
experienced during the RUN (~24 N·m for a 75-kg person),
whereas the external rotation moment during the XOV was
S30 RUN XOV
0.97 (0.01) 0.97 (0.01) 0.96 (0.02)
0.94 (0.04) 0.74 (0.20) 0.88 (0.06)
0.96 (0.02) 0.97 (0.01) 0.96 (0.02)
0.96 (0.03) 0.96 (0.03) 0.95 (0.03)
0.85 (0.18) 0.90 (0.16) 0.87 (0.11)
0.83 (0.08) 0.83 (0.11) 0.85 (0.09)
Medicine & Science in Sports & Exercise姞 1171
FIGURE 3—Flexion/extension moments ap-
plied to the knee during running and cutting
maneuvers (* significant difference com-
pared with the RUN task; P < 0.05).
more than 5 times that of the RUN (~26 N·m for a 75-kg
person).
Knee FE angle. Figures 2– 4 also illustrate the mean
(SD) knee flexion angle during each stage of the stance
phase. The knee flexion angle during PPO was significantly
greater (P ⬍ 0.05) than WA and FPO (WA ⫽ 32.8° ⫾ 6.0°;
PPO ⫽ 45.9° ⫾ 6.3°; FPO ⫽ 23.7° ⫾ 4.6°). The degree of
flexion at the knee was also dependent on the task per-
formed, although the functional significance of these differ-
ences is doubtful. At WA and PPO, the cutting tasks were
performed with ~2° greater knee flexion than the RUN (P ⬍
0.05). This trend was reversed at FPO, where the RUN task
was performed with ~1.5° more knee flexion than the other
tasks (significant for the S60 task only, P ⬍ 0.05).
DISCUSSION
Repeatability of gait data. The FE moments mea-
sured during the RUN in our study were similar to those
reported previously at similar running speeds (17,26). No
external IE or VV moments for running have been reported
previously. The R2 measures for knee joint moments were
higher than those found by Kadaba et al. (8), especially
those measured in IE. Kadaba et al. (8) attributed the vari-
ability in VV and IE moments to inconsistent alignment of
markers. This inconsistency was probably reduced in the
present study by: (a) having only one experimenter to align
the retro-reflective markers on each subject for all trials; (b)
FIGURE 4 —Varus/valgus moments applied
to the knee during running and cutting ma-
neuvers (* significant difference compared
with RUN task; P < 0.01).
1172 Official Journal of the American College of Sports Medicine
using a straight edge to align the thigh marker correctly to
define the thigh plane/segment; and (c) the use of a knee
alignment device contributed to the correct alignment of
knee joint center markers and reliability of VV and IE
measures. The high R2 values for the IE and VV moments
were also due to higher magnitude of these moments during
the cutting tasks compared with those recorded by Kadaba
et al. (8) in walking trials. It was concluded that the gait data
obtained in this study were reliable and repeatable within
and between weeks 2 and 3.
Speed and cutting angle. The speeds at which these
tasks were performed in the laboratory may be less than that
commonly seen in some sporting situations (5–7 m·s-1 in
Rugby League as reported by McLean et al. (14)). However,
this speed ensured that the cutting tasks would be performed
within the limits of the 50-Hz motion analysis system, as
stated previously. In the interest of preserving subjects’
health, the chosen speed was also fast enough to elicit large
VV and IE moments during sidestepping and the crossover
cut with minimal risk of knee ligament injury. An increased
running speed may increase the magnitude of the external
knee joint moments. However, the extent to which the loads
change with running speed is not known. Novacheck (17)
has shown that external flexion moments applied to the knee
actually decrease by ~10% when running speed increased
from 3 to 4 m·s-1. Further research is required to investigate
the effect of increasing running speed on the knee joint loads
during cutting tasks, especially those related to VV and IE.
http://www.acsm-msse.org

Cutting technique may be more important than running
speed or cutting angle in regard to the potential for ligament
injury. In our results, there were technique differences ob-
served during the side stepping tasks (the varus group versus
valgus group), and McLean and colleagues (15) have re-
ported differences in kinematics to perform the same cutting
task. Moreover, athletes routinely perform cutting maneu-
vers without injury at a range of speeds and cutting angles.
So the question should then be asked: what is so different
about a cutting task technique that causes an injury? Impor-
tant technique-related factors associated with increased risk
of noncontact knee ligament injury could include the time
taken to prepare for a maneuver, inappropriate postural
adjustments, or inappropriate muscle activation patterns.
These factors, and others, are currently being investigated in
our laboratory.
It was evident that a compromise between the speed of the
task and the desired cutting angle was required to perform
the cutting maneuvers. Even though the subjects were in-
structed to maintain running speed throughout all the tasks,
they slowed significantly to perform the cutting maneuvers.
This reduction in speed may not only be required to perform
the task, but may also reduce the external loads applied to
the knee.
Technique changes. Five of the subjects performed
the sidestepping tasks with a varus moment applied to the
knee (varus group), whereas the other six subjects experi-
enced a net valgus moment at PPO (valgus group). McLean
et al. (14) found large variations in knee kinematics between
individuals performing sidestepping tasks and concluded
that the level of experience was a major contributing factor
to the degree of consistency. The differences observed in the
VV moments at PPO during the sidestepping tasks in this
study suggest that technique may affect the external load
placed on the joint. Although technique was not measured in
this study, the differences in cutting angles and speed main-
tained throughout the cutting tasks may indicate variations
in the way these maneuvers were performed. The varus
group was able to maintain a greater speed and achieve a
greater angle of sidestep than the valgus group. The varus
group also displayed significantly greater varus loading at
PPO during the XOV task compared to the valgus group.
LOADING AT THE KNEE DURING CUTTING TASKS
FIGURE 5—Internal/external rotation mo-
ments applied to the knee during running
and cutting maneuvers (* significant differ-
ence to RUN task; P < 0.05).
Implications for noncontact knee ligament inju-
ries. Perhaps the most important issue to raise here was
that there was very little difference in the applied flexion
moments at the knee during the different tasks; however,
there were significant differences in the VV and IE loading.
To counter the large flexion loads applied to the knee,
muscles surrounding the joint must apply a large extension
moment that may result in a net anterior force on the tibia
when the knee is near full extension. The combination of
anterior force on the tibia combined with the increased
external VV and IE moments during the cutting tasks may
be a reason for the increased risk of injury to knee joint
ligaments when performing these maneuvers.
The applied VV and IE moments during the cutting
maneuvers were of sufficient magnitude to place large load
on the knee joint ligaments. Piziali et al. (21) performed a
series of loading studies on human cadaver knees and found
that ligament damage occurred within 35– 80 N·m of IE
rotation or 125–210 N·m of VV rotation. The peak IE and
VV moments measured during the cutting tasks of this study
were within the range measured by Piziali et al. (21) and
were coupled with large external flexion loads. Without
adequate muscular support, these loads may well be large
enough to cause injury to knee joint ligaments. However, the
risk of injury depends on many other factors that include:
the magnitude and rate of loading, the combination of loads
applied, and how these loads are apportioned between mus-
cles and other soft tissues surrounding the joint, ligament
strength, and joint geometry. To estimate the loads placed
on ligaments in vivo, a modelling approach similar to that of
Lloyd and Buchanan (10) should be used, taking into ac-
count the individual muscle activation patterns, joint kine-
matics, and joint geometry.
The greatest potential for tension development in the
ACL is during sidestepping at WA and PPO, where the knee
experiences combined loads of anterior tibial force, internal
rotation, and valgus moments (valgus group) and the knee
angle is between 30° and 40° of flexion. However, the
tension on the ACL from these combined loading directions
needs to be qualified, as Markolf et al. (13) only investigated
combined loading from two directions. It is therefore diffi-
cult to predict the force experienced by the ACL under
combined knee loading from all three directions, as
Medicine & Science in Sports & Exercise姞 1173
experienced during the cutting maneuvers. Considerable
ACL loading occurs with combined anterior tibial force and
valgus moments at knee flexion angles between 10° and 40°,
but an internal rotation moment with an anterior tibial force
reduces loading of the ACL at knee flexion angles greater
than 25° compared with anterior tibial force only (13). To
complicate matters further, Markolf et al. (13) also showed
that combined valgus and internal rotation moments in-
creased ACL load compared with the action of each moment
in isolation. Nevertheless, this cadaver work suggests that
the periods of WA and, in particular, PPO in the valgus
group, where the applied internal rotation and valgus mo-
ments are greatest, poses the greatest risk of injury to the
ACL. Further work is required to determine the exact load-
ing response of the ACL to combined anterior tibial force
with valgus and internal rotation moments.
During the crossover cut, even though there were large
external varus and flexion moments that potentially load the
ACL, the risk of ACL injury would be moderated by the
presence of an external rotation moment (13,20). Large
valgus and internal rotation moments applied to the knee at
WA and PPO (in the valgus group) during sidestepping also
have the potential to place high loads on the medial collat-
eral ligament (MCL). Almost 80% of the external valgus
load applied to the knee is supported by the MCL at 25– 30°
knee flexion (7,23), similar knee angles to that observed at
WA. The MCL also supports the large majority of the
applied internal rotation moments between 0° and 45°
(11,23). As the applied valgus and internal rotation mo-
ments increased with the larger cutting angle of the S60
task, so too did the potential for loading of the MCL. Given
this relationship, the varus group may reduce the possible
loading of the MCL compared with the valgus group with-
out compromising the speed and cutting angle achieved.
The large varus loads observed during the crossover cut
have greater potential to damage the lateral collateral ligament
(LCL) (13,20), particularly with knee postures near full exten-
sion (7). Grood et al. (7) showed that the LCL supported close
to 70% of an external varus moment applied to the knee. In
terms of the potential for ligament injury, the varus group may
be at greater risk here compared to the valgus group, owing to
the large varus moments applied at PPO.
Future directions. Further work should concentrate on
the relationships between external knee joint loads and
whole body kinematics during these dynamic functional
tasks. For example, the load experienced at the knee joint
may be related to the position of the foot on the ground with
respect to the body’s center of gravity or the angle of the
lower limb with respect to the ground. Qualitative exami-
REFERENCES
1. ANTONSSON, E. K., and R. W. MANN. The frequency content of gait.
J. Biomech. 18:39 – 47, 1985.
2. BEYNNON, B., J. G. HOWE, M. H. POPE, R. J. JOHNSON, and B. C.
FLEMING. The measurement of anterior cruciate ligament strain in
vivo. Int. Orthop. 16:1–12, 1992.
3. BRANCH, T. P., R. HUNTER, and M. DONATH. Dynamic EMG anal-
ysis of anterior cruciate deficient legs with and without bracing
during cutting. Am. J. Sports Med. 17:35– 41, 1989.
1174 Official Journal of the American College of Sports Medicine
nation of the upper body position suggested that upper body
lean affected the magnitude of the valgus load experienced
at the knee. Such “postural adjustments” need to be quan-
tified, as there may be potential to reduce external loading
at the knee by altering technique.
We are also interested in observing the effect of the
foot-ground interface on the loads placed on the knee joint.
Our subjects performed the running and cutting tasks bare-
foot to negate any effects that may occur from wearing
different shoes. This raises an interesting question: do dif-
ferent shoes alter the external joint loads placed on the knee
joint when performing cutting tasks, particularly in VV and
IE directions? One may expect the VV and IE loads to
increase when wearing shoes compared with barefoot cut-
ting tasks, owing to increased friction between the foot and
the ground. Further research is warranted to investigate the
effects of shoe design on the external VV and IE loads
placed on the knee joint during running and cutting tasks.
SUMMARY
The purpose of this study was to quantify the moments
at the knee in FE, VV, and IE during running, sidestep-
ping, and crossover cutting tasks and to determine the
potential loading of knee ligaments. External flexion
loads were similar among tasks, whereas the external VV
and IE loads placed on the joint increased dramatically
during cutting tasks compared with normal running.
These VV and IE moments are believed to be responsible
for placing knee joint ligaments at a higher risk of injury.
Compared with straight running, the external loads of
flexion, valgus, and internal rotation during sidestepping
have the potential to substantially increase the load ex-
perienced by the ACL and MCL, whereas the combined
flexion and varus loads during the crossover could po-
tentially place large loads on the LCL. These external
loads are of an order of magnitude that could cause
injuries to the knee joint ligaments if the action of the
knee muscles do not adequately support the combined
loads. However, there are other important factors that
may also modulate the risk of injury such as loading rate,
muscle fatigue state, previous ligament injury, mechani-
cal properties of the ligaments, and other anatomical/
geometric factors such as knee and intercondylar notch
width.
Address for correspondence: David Lloyd, Department of Human
Movement & Exercise Science, The University of Western Australia,
Nedlands, Perth, Western Australia 6907; E-mail: dlloyd@cyllene.
uwa.edu.au.
4. CICCOTTI, M. G., R. K. KERLAN, J. PERRY, and M. PINK. An
electromyographic analysis of the knee during functional activi-
ties: I. The normal profile. Am. J. Sports Med. 22:645– 650, 1994.
5. CROSS, M. J., N. J. GIBBS, and G. J. BRYANT. An analysis of the
sidestep cutting manoeuvre. Am. J. Sports Med. 17:363–366, 1989.
6. DAVIS, R. B., S. OUNPUU, D. TYBURSKI, and J. R. GAGE. A gait
analysis data collection and reduction technique. Hum. Mov. Sci.
10:575–587, 1991.
http://www.acsm-msse.org
 7. GROOD, E. S., F. R. NOYES, D. L. BUTLER, and W. J. SUNTAY.
Ligamentous and capsular restraints preventing straight medial
and lateral laxity in intact human cadaver knees. J. Bone Joint
Surg. (Am.) 63:1257–1269, 1981.
8. KADABA, M. P., H. K. RAMAKRISHNAN, M. E. WOOTTEN, J. GAINEY,
G. GORTON, and G. V. COCHRAN. Repeatability of kinematic, ki-
netic, and electromyographic data in normal adult gait. J. Orthop.
Res. 7:849 – 860, 1989.
9. KADABA, M. P., H. K. RAMAKRISHNAN, and M. E. WOOTTEN. Mea-
surement of lower extremity kinematics during level walking.
J. Orthop. Res. 8:383–392, 1990.
10. LLOYD, D. G., and T. S. BUCHANAN. A model of load sharing
between muscles and soft tissues at the human knee during static
tasks. J. Biomech. Eng. 118:367–376, 1996.
11. MARKOLF, K. L., J. S. MENSCH, and H. C. AMSTUTZ. Stiffness and
laxity of the knee: the contributions of the supporting structures.
J. Bone Joint Surg. (Am.) 58A:583–593, 1976.
12. MARKOLF, K. L., D. C. WASCHER, and G. A. FINERMAN. Direct in
vitro measurement of forces in the cruciate ligaments. Part II: The
effect of section of the posterolateral structures. J. Bone Joint
Surg. (Am.) 75:387–394, 1993.
13. MARKOLF, K. L., D. M. BURCHFIELD, M. M. SHAPIRO, M. F. SHEP-
ARD, G. A. FINERMAN, and J. L. SLAUTERBECK. Combined knee
loading states that generate high anterior cruciate ligament forces.
J. Orthop. Res. 13:930 –935, 1995.
14. MCLEAN, S. G., P. T. MYERS, R. J. NEAL, and M. R. WALTERS. A
quantitative analysis of knee joint kinematics during the sidestep
cutting maneuver: implications for non-contact anterior cruciate
ligament injury. Bull. Hosp. Joint Dis. 57:30 –38, 1998.
15. MCLEAN, S. G., R. J. NEAL, P. T. MYERS, and M. R. WALTERS. Knee
joint kinematics during the sidestep cutting maneuver: potential
for injury in women. Med. Sci. Sports Exerc. 31:959 –968, 1999.
LOADING AT THE KNEE DURING CUTTING TASKS
16. NEPTUNE, R. R., I. C. WRIGHT, and A. J. VAN DEN BOGERT. Muscle
coordination and function during cutting movements. Med. Sci.
Sports Exerc. 31:294 –301, 1999.
17. NOVACHECK, T. F. The biomechanics of running. Gait & Posture
7:77–95, 1998.
18. OUNPUU, S. The biomechanics of walking and running. Clin.
Sports Med. 13:843– 863, 1994.
19. PANDY, M. G., and K. B. SHELBURNE. Dependence of cruciate-
ligament loading on muscle forces and external load. J. Biomech.
30:1015–1024, 1997.
20. PIZIALI, R. L., J. RASTEGAR, D. A. NAGEL, and D. J. SCHURMAN. The
contribution of the cruciate ligaments to the load-displacement
characteristics of the human knee joint. J. Biomech. Eng. 102:
277–283, 1980.
21. PIZIALI, R. L., D. A. NAGEL, T. KOOGLE, and R. WHALEN. Knee and
tibia strength in snow skiing. In R. J. Johnson, W. Hauser, and M.
Magi (Eds.). In Ski Trauma and Skiing Safety IV. Munich: TUV
Publication Series, 1982, pp. 24 –31.
22. RYDER, S. H., R. J. JOHNSON, B. D. BEYNNON, and C. F. ETTLINGER.
Prevention of ACL injuries. J. Sport Rehabil. 6:80 –96, 1997.
23. SEERING, W. P., R. L. PIZIALI, D. A. NAGEL, and D. J. SCHURMAN.
The function of the primary ligaments of the knee in varus-valgus
and axial rotation. J. Biomech. 13:785–794, 1980.
24. WASCHER, D. C., K. L. MARKOLF, M. S. SHAPIRO, and G. A.
FINERMAN. Direct in vitro measurement of forces in the cruciate
ligaments. Part I: The effect of multiplane loading in the intact
knee. J. Bone Joint Surg. (Am.) 75:377–386, 1993.
25. WHITING, W. C., and R. F. ZERNICKE. Biomechanics of Muscu-
loskeletal Injury. Champaign, IL: Human Kinetics, 1998, p.
177.
26. WINTER, D. A. Moments of force and mechanical power in jog-
ging. J. Biomech. 16:91–97, 1983.
Medicine & Science in Sports & Exercise姞 1175