Clinical Biomechanics 25 (2010) 213221

Contents lists available at ScienceDirect

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

Cruciate ligament tensile forces during the forward and side lunge
Rafael F. Escamilla a,i,*, Naiquan Zheng b, Toran D. MacLeod c, Rodney Imamura d, W. Brent Edwards e,
Alan Hreljac d, Glenn S. Fleisig f, Kevin E. Wilk g, Claude T. Moorman III h, Lonnie Paulos i,
James R. Andrews f,i
a

Department of Physical Therapy, California State University, Sacramento, CA, USA

The Center for Biomedical Engineering, Department of Mechanical Engineering and Engineering Science, University of North Carolina at Charlotte, NC, USA
Department of Physical Therapy, Center for Biomedical Engineering Research, University of Delaware, Newark, DE, USA
d
Kinesiology and Health Science Department, California State University, Sacramento, USA
e
Department of Kinesiology and Nutrition, University of Illinois at Chicago, IL, USA
f
American Sports Medicine Institute, Birmingham, AL, USA
g
Champion Sports Medicine, Birmingham, AL, USA
h
Duke University Medical Center, Durham, NC, USA
i
Andrews-Paulos Research and Education Institute, Gulf Breeze, FL, USA
b
c

a r t i c l e

i n f o

Article history:
Received 21 August 2008
Accepted 3 November 2009

Keywords:
Anterior cruciate ligament
Posterior cruciate ligament
Knee kinetics
Rehabilitation
Closed chain

a b s t r a c t
Background: Although weight bearing lunge exercises are frequently employed during anterior cruciate
ligament and posterior cruciate ligament rehabilitation, cruciate ligament tensile forces are currently
unknown while performing forward and side lunge exercises with and without a stride.
Methods: Eighteen subjects used their 12 repetition maximum weight while performing a forward lunge
and side lunge with and without a stride. A motion analysis system and biomechanical model were used
to estimate cruciate ligament forces during lunging as a function of 090 knee angles.
Findings: Comparing the forward lunge to the side lunge across stride variations, mean posterior cruciate
ligament forces ranged between 205 and 765 N and were signicantly greater (P < 0.0025) in the forward
lunge long at 40, 50, 60, 70, and 80 knee angles of the descent phase and at 80, 70, 60 knee angles
of the ascent phase. There were no signicant differences (P < 0.0025) in mean posterior cruciate ligament forces between with and without stride differences across lunging variations. There were no anterior cruciate ligament forces quantied while performing forward and side lunge exercises.
Interpretation: Clinicians should be cautious in prescribing forward and side lunge exercises during early
phases of posterior cruciate ligament rehabilitation due to relatively high posterior cruciate ligament
forces that are generated, especially during the forward lunge at knee angles between 40 and 90 knee
angles. Both the forward and side lunges appear appropriate during all phases of anterior cruciate ligament rehabilitation. Understanding how forward and side lunging affect cruciate ligament loading over
varying knee angles may help clinicians better prescribe lunging exercises in a safe manner during anterior cruciate ligament and posterior cruciate ligament rehabilitation.
2009 Elsevier Ltd. All rights reserved.

1. Introduction
Closed chain weight bearing exercises, such as the squat, leg
press, and lunge, and activities of daily living, such as walking,
are commonly used in rehabilitation settings, such as after anterior
cruciate ligament (ACL) or posterior cruciate ligament (PCL) reconstruction surgery (Bynum et al., 1995; Escamilla et al., 1998; Shelburne et al., 2004, 2005; Yack et al., 1993). Both the squat and leg
press have been shown to generate very low or negligible ACL force
during the squat and leg press, with peak ACL forces typically less
* Corresponding author. Address: Department of Physical Therapy, California
State University, 6000 J Street, Sacramento, CA 95819-6020, USA.
E-mail address: rescamil@csus.edu (R.F. Escamilla).
0268-0033/$ - see front matter 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.clinbiomech.2009.11.003

than 60 N (Escamilla et al., 1998, 2009; Shelburne and Pandy,

2002; Toutoungi et al., 2000). During normal walking, peak ACL
forces are typically between 150 and 300 N (Morrison, 1970; Shelburne et al., 2004, 2005), while during drop-landings ACL forces are
approximately 300 N (Pum et al., 2004). Therefore, performing
closed chain weight bearing exercises and activities, such as walking, squatting, the leg press, and jumping/landing, typically generates peak loads in the ACL that are only 1015% of the maximum
strength of the normal healthy ACL, which has been reported to
be approximately 2200 N (Woo et al., 1991).
Compared to ACL forces during closed chain weight bearing exercises, peak PCL forces are higher. For example, peak PCL forces are
400500 N during the one leg squat (Escamilla et al., 2009), 700
800 N during the wall squat (Escamilla et al., 2009), and

214

R.F. Escamilla et al. / Clinical Biomechanics 25 (2010) 213221

18001900 N during the barbell squat and leg press (Escamilla et al.,
1998). However, because maximum strength of the normal healthy
PCL is approximately twice as strong compared to the normal
healthy ACL (Race and Amis, 1994), PCL loading is also relatively
low during these types of closed chain weight bearing exercises.
There are several factors that can affect ACL and PCL forces
while performing exercises and functional activities, such as the
magnitude and direction of the ground reaction force, muscle
forces (primarily from the quadriceps, hamstrings, and gastrocnemius), tibiofemoral compressive loading, rotation of the tibia with
respect to the femur, and knee angle.(Escamilla et al., 1998, 2009;
Shelburne et al., 2004, 2005) These exercises can all be performed
with technique variations, which may affect ACL and PCL loading.
For example, when employing the same relative intensity (12 repetition maximum), PCL forces varied considerably between the
barbell squat, wall squat, and one leg squat (Escamilla et al.,
1998, 2009). ACL forces were also signicantly different between
the one leg squat and the wall squat (Escamilla et al., 2009).
Although the effects of exercise technique variations on cruciate
ligament loading have been examined while performing the squat
and leg press (Escamilla et al., 1998, 2001), there are no studies
that have examined the effects of technique variations on cruciate
ligament loading while performing lunging exercises. However, it
has been demonstrated that hamstring activity increased when
the forward lunge was performed with a more forward inclined
trunk compared to a more erect trunk, which may increase PCL
loading and decrease ACL loading (Farrokhi et al., 2008).
During knee rehabilitation the lunge is often prescribed using
multiple techniques, such as performing a forward or side lunge,
or lunging with or without a stride. Lunging forward with a stride
involves stepping forward from the upright position with the lead
leg and pushing back to the starting position. Lunging forward
without a stride involves positioning one foot ahead of the other
and lunging up and down while keeping both feet stationary. All
of these different lunging technique variations may affect cruciate
ligament loading.
Understanding how cruciate ligaments are loaded differently
among technique variations of the forward and side lunges may allow clinicians to prescribe safer and more effective knee treatment
to patients during ACL or PCL rehabilitation. For example, if ACL
loading is greater during the side lunge compared to the forward
lunge, the forward lunge may be more appropriate for the patient
if the clinicians immediate goal for their patient was to minimize
ACL loading. Similarly, if PCL loading is greater with a stride than
without a stride, lunging without a stride may be more appropriate
for the patient if the clinicians immediate goal for their patient
was to minimize PCL loading.
The purpose of this study was to compare cruciate ligament
tensile forces while performing forward and side lunges with and
without a stride. It was hypothesized that PCL tensile forces would
be greater during the forward lunge compared to the side lunge,
and greater with a stride compared to without a stride. It was
hypothesized that this greater PCL loading would occur because
of greater quadriceps force generated during the forward lunge
compared to the side lunge. It was also hypothesized that ACL tensile forces would be minimal (<150 N) for all lunge variations.
Quadriceps and hamstrings muscle force magnitudes will also be
described to help better understand ACL and PCL force magnitudes.

age (SD) age, mass, and height of 29 (7) y, 77 (9) kg, and 177 (6)
cm, respectively, for males, and 25 (2) y, 60 (4) kg and 164 (6)
cm, respectively, for females. All subjects were required to be able
to perform all exercises pain-free and with proper form and technique for 12 consecutive repetitions using their 12 repetition maximum (12 RM) weight for both the forward lunge and side lunge.
The mean (SD) mass of both dumbbells during the forward lunge
was 49 (11) kg for males and 32 (8) kg for females. On the average
(SD), the dumbbell mass used during the side lunge was 34 (9) kg
for males and 20 (5) kg for females. The average (SD) stride length
(measured from left toe to right heel) during the forward lunge of
89 (4) cm for males and 79 (6) cm for females. The average (SD)
stride length (measured from inside of left heel to inside of right
heel) during the side lunge was 94 (5) cm for males and 83 (3)
cm for females.
To control the electromyographic (EMG) signal quality, the current study was limited to males and females that had average or
below average body fat, which was assessed by Baseline skinfold
calipers (Model 68900, Country Technology, Inc., Gays Mill, WI,
USA) and appropriate regression equations and body fat standards
set by the American College of Sports Medicine (Balady et al.,
2000). Average (SD) body fat was 12 (4)% for males and 18 (1)%
for females. All subjects provided written informed consent in
accordance with the Institutional Review Board at California State
University, Sacramento, USA, which approved the research conducted and informed consent form.
2.2. Exercise description
2.2.1. Forward Lunge
Each subject performed the forward lunge (Fig. 1a) both with
and without a stride. The starting and ending positions for the forward lunge with a stride were the same, which involved standing
upright with both feet together and the knees fully extended (full
knee extension = 0 knee exion angle). From the starting position
for the forward lunge with a stride, the subject held a dumbbell
weight in each hand and lunged forward with the right leg toward
a force platform at ground level. At right foot contact the right knee
slowly exed until maximum right knee exion of 90100 was
obtained as the left knee made contact with the ground. From this
ending position the subject immediately pushed backward off the
force platform and returned to the starting position. A metronome
was used to ensure that the knee exed and extended at approximately 45/s. Each subject was instructed to use as long a stride

2. Methods
2.1. Subjects
Eighteen healthy individuals (9 males and 9 females) without a
history of cruciate ligament pathology participated with an aver-

Fig. 1a. Forward lunge.

R.F. Escamilla et al. / Clinical Biomechanics 25 (2010) 213221

length as was comfortable. A tester ensured that the stride was

long enough so that at the lowest position of the lunge the stride
knee was maintained over the stride foot without translating forward beyond the toes. During the forward lunge, a longer stride
length is commonly preferred by individuals compared to a shorter
stride length to ensure that the stride knee does not progress beyond the toes at the lowest position of the forward lunge
(Fig. 1a). The forward lunge without stride was performed the
same as the forward lunge with stride with the exception that both
feet remained stationary throughout each repetition during the
forward lunge without stride. That is, from the lowest position of
the forward lunge shown in Fig. 1a, the subject simply fully extended both knees, and then exed both knees back to return to
the lowest position of the forward lunge shown in Fig. 1a. During
the forward lunge with and without a stride maximum forward
trunk tilt (which occurred near maximum lead knee exion) was
approximately 1020 (relative to a vertical axis) for all subjects.
2.2.2. Side lunge
Each subject performed the side lunge (Fig. 1b) both with and
without a stride. The starting and ending positions for the side
lunge with a stride were both the same, which involved standing
upright with both feet together. From the starting position for
the side lunge with a stride, the subject held a single dumbbell
weight down between the legs and lunged sideways with the left
knee remaining fully extended and the right leg moving toward a
force platform at ground level. At right foot contact the right foot
was turned out approximately 3045 relative to the left foot (subjects preference) and the right knee exed slowly at approximately 45/s until the right knee exed approximately 90100
(Fig. 1b) as the left knee remained fully extended. From this ending
position the subject then pushed backward off the force platform
and returned to the starting position. A metronome was used to
ensure that the knee exed and extended at approximately 45/s.
Each subject was instructed to use as long a stride length as was
comfortable. A tester ensured that the stride was long enough so
that at the lowest position of the lunge the stride knee was maintained over the stride foot without translating forward beyond the
toes. Using a longer stride length compared to a shorter stride
length during the side lunge is commonly preferred by individuals
as it allows the left knee to remain straight and the right knee to
ex approximately 90100 and remain over the foot. The side
lunge without stride was performed the same as the side lunge
with stride with the exception that both feet remained stationary

Fig. 1b. Side lunge.

215

throughout each repetition during the side lunge without stride.

That is, from the lowest position of the side lunge shown in
Fig. 1b, the subject simply fully extended the right knee (left knee
was already extended), and then returned back to the lowest position of the side lunge by exing the right knee). During the side
lunge with and without a stride maximum forward trunk tilt
(which occurred near maximum lead knee exion) was approximately 1020 (relative to a vertical axis) for all subjects, similar
to forward trunk tilt during the forward lunge. Also, for both the
side lunge and forward lunge the right foot pointed in the same
direction and the right lower extremity moved in the same plane.

2.3. Data collection

Each subject came in for a pre-test one week prior to the testing
session. At that time the experimental protocol was reviewed and
the subject was given the opportunity to ask questions. In addition,
each subjects 12 RM was determined for the forward and side
lunges by utilizing the most weight they could lift for 12 consecutive repetitions. Moreover, each subjects stride length, as previously dened, was measured for the forward and side lunges.
Blue Sensor (Ambu Inc., Linthicum, MD, USA) disposable surface
electrodes (type M-00-S) were used to collect EMG data. These oval
shaped electrodes (22 mm wide and 30 mm long) were placed in a
bipolar electrode conguration along the longitudinal axis of each
muscle, with a center-to-center distance of approximately 3 cm
between electrodes. Prior to positioning the electrodes over each
muscle, the skin was prepared by shaving, abrading, and cleaning
with isopropyl alcohol wipes to reduce skin impedance. As previously described (Basmajian and Blumenstein, 1980), electrode
pairs were then placed on the subjects right side for the following
muscles: (1) rectus femoris; (2) vastus lateralis; (3) vastus medialis; (4) medial hamstrings (semitendinosis and semitendinosus);
(5) lateral hamstrings (biceps femoris); and (6) gastrocnemius.
Spheres (3.8 cm in diameter) were attached to adhesives and positioned over the following bony landmarks: (a) third metatarsal
head of the right foot; (b) medial and lateral malleoli of the right
leg; (c) upper edges of the medial and lateral tibial plateaus of
the right knee; (d) posterosuperior greater trochanters of the left
and right femurs; and (e) lateral acromion of the right shoulder.
Once the electrodes and spheres were positioned, the subject
warmed up and practiced the exercises as needed, and data collection commenced. A 6 camera Peak Performance motion analysis
system (Vicon-Peak Performance Technologies, Inc., Englewood,
CO) was used to collect 60 Hz video data. Force data were collected
at 960 Hz using a force platform (Model OR6-6-2000, Advanced
Mechanical Technologies, Inc.). EMG data were collected at
960 Hz using a Noraxon Myosystem unit (Noraxon USA, Inc.,
Scottsdale, AZ). The EMG amplier had a bandwidth frequency of
10500 Hz and an input impedance of 20,000 kX, with a common-mode rejection ratio of 130 dB. Video, EMG, and force data
were electronically synchronized and simultaneously collected as
each subject performed in a randomized manner one set of three
continuous repetitions (trials) during the forward and side lunges
with and without a stride.
Prior to the exercise trials, EMG data were collected during
maximum voluntary isometric contractions (MVIC) to normalize
the EMG data collected during each exercise (Escamilla et al.,
1998). The MVIC for the rectus femoris, vastus lateralis, and vastus
medialis were collected in a seated position at 90 knee and hip
exion with a maximum effort knee extension. The MVIC for the
lateral and medial hamstrings were collected in a seated position
at 90 knee and hip exion with a maximum effort knee exion.
MVIC for the gastrocnemius was collected during a maximum effort standing one leg toe raise with the ankle positioned approxi-

216

R.F. Escamilla et al. / Clinical Biomechanics 25 (2010) 213221

mately halfway between neutral and full plantar exion. Two 5 s

trials were collected for each MVIC.
2.4. Data reduction
Video images for each reective marker were tracked and digitized in three-dimensional space with Peak Performance software.
Ankle, knee, and hip joint centers were mathematically determined
using the external markers and appropriate equations previously
described (Escamilla et al., 1998). Testing of the accuracy of the calibration system resulted in markers that could be located in threedimensional space with an error less than 7 mm. The raw position
data were smoothed with a double-pass fourth order Butterworth
low-pass lter with a cut-off frequency of 6 Hz (Escamilla et al.,
1998). Joint angles, linear and angular velocities, and linear and
angular accelerations were calculated utilizing appropriate kinematic equations (Escamilla et al., 1998).
Raw EMG signals were full-waved rectied and smoothed with
a 10 ms moving average window (linear enveloped) throughout
the knee range of motion for each repetition. These EMG data were
then normalized for each muscle and expressed as a percentage of
each subjects highest corresponding MVIC trial. The MVIC trials
were calculated using the highest EMG signal over a 1 s time interval throughout the 5 s MVIC. Normalized EMG data for the three
repetitions (trials) were then averaged at corresponding knee angles between 0 and 90, and were used in the biomechanical model described below.
2.5. Biomechanical model
As previously described (Escamilla et al., 1998; Zheng et al.,
1998), a biomechanical model of the knee (Fig. 2) was used to continuously estimate cruciate ligament forces throughout a 90 knee
range of motion during the knee exing (lunge descent) phase (0
90) and the knee extending (lunge ascent) phase (900) of the
lunge exercises. Resultant force and torque equilibrium equations
were calculated using inverse dynamics and the biomechanical
knee model (Escamilla et al., 1998; Zheng et al., 1998). Anteroposterior shear forces in the knee were calculated and adjusted to ligament orientations to estimate ACL or PCL forces (Herzog and Read,
1993). Moment arms of muscle forces and angles for the line of ac-

tion for the muscles and cruciate ligaments were expressed as

polynomial functions of knee angle using data from Herzog and
Read (Herzog and Read, 1993). Knee torques from cruciate and collateral ligament forces and bony contact were assumed to be negligible, as were forces and torques out of the sagittal plane.
Quadriceps, hamstrings, and gastrocnemius muscle forces were
estimated as previously described (Escamilla et al., 1998; Zheng
et al., 1998). Because the accuracy of estimating muscle forces depends on accurate estimations of a muscles physiological crosssectional area (PCSA), maximum voluntary contraction force per
unit PCSA, and the EMGforce relationship, resultant force and torque equilibrium equations may not be satised. Therefore, each
muscle force Fm(i) was modied by the following equation at each
knee angle:

F mi ci kli kv i Ai rmi EMGi =MVICi ;

where Ai was PCSA of the ith muscle, rm(i) was MVIC force per unit
PCSA of the ith muscle, EMGi and MVICi were EMG window averages of the ith muscle EMG during exercise and MVIC trials, ci
was a weight factor (values given below) adjusted in a computer
optimization program to minimize the difference between the
resultant torque from the inverse dynamics (Tres) and the resultant
torque calculation from the biomechanical model (Tmi) (Fig. 2), kli
represented each muscles forcelength relationship as function of
hip and knee angles (based on muscle length, ber length, sarcomere length, pennation angle, and cross-sectional area) (Wickiewicz et al., 1983), and kvi represented each muscles force
velocity relationship based on a Hill-type model for eccentric and
concentric muscle actions (Epstein and Herzog, 1998; Zajac,
1989). The specic mathematical calculations and values used for
ci, kvi, Ai, and rm(i) are shown in Appendix A.
2.6. Data analysis
To determine signicant differences in cruciate ligament forces
between the two lunge variations (forward lunge and side lunge)
and the two stride variations (with stride and without stride), cruciate ligament forces were statistically analyzed every 10 during
the 090 knee exing (descent) phase and the 900 knee extending (ascent) phase using a two factor (exercise variations and stride
variations) repeated measures Analysis of Variance. To minimize
the probability of type I errors secondary to the use of a separate
ANOVA for each knee angle, a Bonferroni adjustment was performed with the level of signicance established at 0.0025 (0.05/
20 knee angles).
3. Results

Fig. 2. Computer optimization with input from measured knee torque from inverse
dynamics and predicted knee torque from muscle model, where TK = resultant knee
torque, FK = resultant knee force, I = moment of inertia about leg center of mass,
a = angular acceleration of leg, m = mass of leg, a = linear acceleration of leg, g is
gravitation constant 9.80 m/s2, Fext = external force acting on foot, Text = external
torque acting on foot, FQ = quadriceps force, FP = patellar tendon force, FH = hamstrings force, and FG = gastrocnemius force. Note: to simplify the drawing the equal
and opposite forces and torques acting on the distal leg and proximal ankle are not
shown.

Cruciate ligament force magnitudes and patterns are shown in

Table 1 and Figs. 36. Comparing the forward lunge to the side
lunge across stride variations (Table 1), mean PCL forces ranged between 205 and 765 N and were signicantly greater (P < 0.0025) in
the forward lunge long at 40, 50, 60, 70, and 80 knee angles of
the descent phase and at 80, 70, 60 knee angles of the ascent
phase. There were no signicant differences (P < 0.0025) in mean
PCL forces between with and without stride differences across step
length variations (Table 1). There were no signicant interactions
between levels of exercise (forward lunge and side lunge) and
stride (with stride and without stride). There were no ACL forces
quantied while performing forward and side lunge exercises.
Descriptive data of mean quadriceps and hamstrings force values during forward and side lunge exercises are shown in Table 2.
For both the forward lunge and side lunge, quadriceps force ranged
between approximately 50650 N and generally increased with
knee exion, while hamstring force ranged between approximately

217

R.F. Escamilla et al. / Clinical Biomechanics 25 (2010) 213221

Table 1
Mean (SD) cruciate ligament force (N) values between exercise variations (forward and side lunge) and between stride variations (with and without stride). Anterior cruciate
ligament forces are negative values and posterior cruciate ligament forces are positive values.
Exercise variations

Stride variations

Knee angles

Forward lunge

Side lunge

P-value

With stride

Without stride

P-value

Descent phase
0
10
20
30
40
50
60
70
80
90

349
396
420
387
383
422
474
521
570
591

(202)
(177)
(205)
(207)
(185)
(155)
(141)
(173)
(182)
(218)

205
267
280
256
239
213
266
314
394
459

(95)
(132)
(134)
(166)
(151)
(96)
(141)
(140)
(127)
(124)

0.151
0.196
0.107
0.028
0.002*
<0.001*
<0.001*
<0.001*
0.002*
0.023

216
267
303
333
346
352
407
439
488
510

(158)
(171)
(182)
(217)
(228)
(173)
(167)
(171)
(179)
(141)

339
392
396
329
291
298
348
412
491
552

(171)
(146)
(191)
(186)
(169)
(160)
(179)
(207)
(229)
(234)

0.076
0.036
0.096
0.996
0.129
0.161
0.117
0.384
0.626
0.710

Ascent phase
90
80
70
60
50
40
30
20
10
0

682
740
765
744
706
676
657
580
488
412

(275)
(247)
(220)
(233)
(264)
(279)
(259)
(268)
(212)
(185)

521
523
567
611
641
617
556
455
351
286

(138)
(158)
(184)
(240)
(326)
(325)
(288)
(226)
(167)
(127)

0.021
0.002*
<0.001*
0.002*
0.127
0.171
0.014
0.008
0.025
0.766

581
636
700
733
712
654
610
467
353
365

(185)
(210)
(224)
(260)
(299)
(283)
(271)
(249)
(157)
(161)

634
652
652
645
651
648
615
570
478
359

(276)
(258)
(228)
(223)
(287)
(315)
(281)
(255)
(218)
(185)

0.616
0.653
0.104
0.043
0.063
0.401
0.522
0.475
0.390
0.235

Note: The mean values given for the exercise variations (forward lunge and side lunge) were collapsed across the two stride variations (with stride and without stride), while
the mean values given for the two stride variations were collapsed across the two exercise variations. The P-values shown for exercise variations and stride variations
represent the main effects of the ANOVA.
*
Signicant difference (P < 0.0025) between exercise variations or stride variations.

Fig. 3. Mean (SD) PCL tensile force during forward and side lunges without stride.

50150 N and remained relatively constant throughout the descent phase and throughout the ascent phase. At each knee angle,
quadriceps and hamstrings forces were generally greater during
the ascent phase compared to the descent phase.

4. Discussion
Our results demonstrate that when the goal is to minimize ACL
loading, such as during the early phases of ACL rehabilitation after
ACL reconstruction, both forward and side lunge exercises may be
appropriate based on the absence of ACL forces while performing

these exercises. In contrast, when the goal is to minimize PCL loading, such as during the early phases of PCL rehabilitation after PCL
reconstruction, both forward and side lunge exercises should be
used cautiously due to the relatively high PCL forces (near bodyweight in magnitudes) that were generated while performing
these exercises.
Like the current study, several studies have also reported either
the absence of ACL loading or minimal ACL loading while performing lunging and other similar weight bearing exercises. Stuart et al.
(1996) reported tibial posterior shear forces (PCL loading), but not
anterior shear forces (ACL loading) throughout the knee range of
motion while performing a forward lunge exercise using a 50 N

218

R.F. Escamilla et al. / Clinical Biomechanics 25 (2010) 213221

Fig. 4. Mean (SD) PCL tensile force during forward and side lunges with stride.

Fig. 5. Mean (SD) PCL tensile force during forward lunge with and without stride.

barbell, which support the results of the current study. According

to Butler et al. (1980)), the ACL provides 86% of the total resistance
to anterior drawer, caused by anterior shear forces, and the PCL
provides approximately 95% of the total restraining force to posterior drawer, caused by posterior shear forces. Escamilla et al. (1998,
2001) also reported PCL loading and the absence of ACL loading
throughout the knee range of motion during the barbell squat
and leg press using a 12 repetition maximum load.
When ACL forces have been reported during closed chain
weight bearing exercises, these forces are always relatively low
with respect to the maximum strength of the normal healthy
ACL, which is approximately 2200 N (Woo et al., 1991). Low or negligible ACL forces were generated during the leg press and barbell
or dumbbell squatting exercises, with peak ACL forces typically
less than 60 N (Escamilla et al., 1998, 2009; Shelburne and Pandy,
2002; Toutoungi et al., 2000). The peak ACL forces of approximately 150300 N during normal walking (Morrison, 1970; Shelburne et al., 2004, 2005) and approximately 300 N during drop-

landings (Pum et al., 2004) are only 1015% of the maximum

strength of the ACL. Therefore, walking and drop-landings may
be appropriate in the healthy ACL and in the later stages of ACL
rehabilitation, assuming that normal landing mechanics are employed during drop-landings. Unfortunately, especially in female
athletes, improper landing techniques, such as inadequate knee
exion combined with knee valgus and rotation, can result in an
ACL injury (Ford et al., 2003).
While ACL forces in the aforementioned studies were estimated
through a variety of modeling techniques, there have been in vivo
studies performed during closed chain weight bearing exercises
(Heijne et al., 2004), and these studies have also demonstrated relatively small magnitude (estimated to be less than 150 N based on
the nding that a 150 N Lachman Test produced 3.7% strain at
30 knee exion angle) ACL loading at lower knee angles (less than
60) during lunging and squatting type exercises. During a bodyweight lunge with a stride, Heijne et al. (2004) reported a mean
ACL strain in vivo of approximately 1% or less (approximately

219

R.F. Escamilla et al. / Clinical Biomechanics 25 (2010) 213221

Fig. 6. Mean (SD) PCL tensile force during side lunge with and without stride.

Table 2
Mean (SD) quadriceps and hamstrings force values during forward and side lunge exercises.
Quadriceps (N)

Hamstrings (N)

Exercise variations

Stride variations

Exercise variations

Stride variations

Knee angles

Forward lunge

Side lunge

With stride

Without stride

Forward lunge

Side lunge

With stride

Without stride

Descent phase
0
10
20
30
40
50
60
70
80
90

87 (84)
111 (67)
131 (68)
179 (80)
237 (117)
326 (163)
435 (186)
551 (204)
660 (157)
540 (172)

56 (45)
122 (92)
161 (113)
180 (123)
231 (141)
297 (145)
413 (169)
535 (187)
622 (195)
542 (145)

132
152
197
223
254
321
429
546
597
576

(78)
(116)
(132)
(127)
(147)
(159)
(181)
(198)
(204)
(176)

71 (52)
81 (44)
95 (49)
136 (76)
203 (111)
302 (148)
419 (174)
540 (193)
586 (148)
506 (140)

47
64
66
69
67
70
71
67
60
57

53
55
56
55
56
56
56
46
39
43

35
51
56
61
64
69
75
67
54
53

66
67
66
63
59
57
52
46
44
47

Ascent phase
90
80
70
60
50
40
30
20
10
0

500
564
646
652
486
409
336
268
206
136

499 (65)
550 (173)
639 (217)
627 (219)
533 (180)
406 (141)
311 (121)
216 (96)
153 (84)
75 (51)

409
509
609
617
561
462
385
298
235
128

(112)
(186)
(224)
(244)
(228)
(192)
(174)
(147)
(137)
(97)

390 (73)
475 (144)
546 (167)
542 (167)
458 (141)
352 (114)
262 (92)
185 (68)
124 (52)
83 (44)

101
108
120
128
134
139
143
140
142
121

(120)
(156)
(175)
(192)
(188)
(165)
(144)
(119)
(105)
(90)

40 N or less) at knee angles less than 60 (no ACL strain at knee

exion angles greater than 60), and a peak ACL strain of 1.8%
(approximately 75 N) between a 0 and 30 knee angle range.
Moreover, Beynnon et al. (1997) reported a peak in vivo ACL strain
of approximately 4% (approximately 150 N) at knee angles between 0 and 60 during squatting with and without a low resistance sport cord, and no ACL strain at knee exion angles greater
than 60. These in vivo studies are unique in that they calculated
ACL strain by direct measurement using force sensors within the
ACL. The subjects in Heijne et al. (2004) and Beynnon et al.
(1997) were patients that had force sensors implanted within the
anteromedial bundle of a healthy ACL during arthroscopic surgery
to repair damaged knee structures (partial miniscectomies; capsule, and patellofemoral joint debridement). Immediately after surgery these patients were asked to perform a variety of exercises,
including the forward lunge, and the strain within the anterome-

(21)
(38)
(40)
(39)
(39)
(36)
(41)
(43)
(37)
(33)
(48)
(48)
(58)
(63)
(64)
(71)
(74)
(75)
(80)
(84)

(38)
(37)
(38)
(34)
(32)
(37)
(43)
(34)
(23)
(31)

73 (36)
88 (41)
108 (50)
125 (59)
136 (64)
141 (73)
138 (74)
129 (76)
115 (70)
108 (80)

(31)
(44)
(47)
(45)
(43)
(44)
(54)
(50)
(35)
(34)

92 (49)
100 (47)
122 (59)
139 (70)
152 (74)
158 (85)
158 (90)
147 (92)
136 (95)
109 (97)

(28)
(31)
(31)
(28)
(27)
(29)
(29)
(27)
(25)
(29)

81 (35)
95 (42)
107 (49)
113 (52)
119 (55)
122 (59)
122 (58)
123 (59)
122 (55)
119 (49)

dial bundle of the ACL was measured and referenced to an instrumented Lachman Test. Because these in vivo studies by Heijne
et al. (2004)) and Beynnon et al. (1997) did not measure PCL strain
during lunging and squatting studies, PCL loads cannot be compared between these studies and the current study.
One limitation to the aforementioned squat and lunge in vivo
studies is that it is unknown what lunging and squatting techniques
were employed while performing these exercises. Both ACL and PCL
forces while performing lunging and squatting exercises are dependent on specic exercise technique. For example, in Beynnon et al.
(1997) it appears the subjects may have squatted using a more upright trunk position with relatively little forward trunk tilt, which
suggests these subjects may use their quadriceps to a greater extent
than their hamstrings (Ohkoshi et al., 1991). This is important because hamstrings force has been shown to unload the ACL and load
the PCL during the weight bearing squat exercise (Escamilla et al.,

220

R.F. Escamilla et al. / Clinical Biomechanics 25 (2010) 213221

1998; More et al., 1993; Ohkoshi et al., 1991). Ohkoshi et al. (1991)
reported no ACL loading at all knee angles tested (15, 30, 60, and
90) while maintaining a squat position with the trunk tilted forward, with 30 or more forward trunk tilt being optimal for eliminating or minimizing ACL strain throughout the knee range of motion.
In addition, Farrokhi et al. (2008) demonstrated that compared to
performing a forward lunge with a relatively erect trunk (similar
to the forward trunk tilt in the current study), performing the lunge
with the trunk tilted forward approximately 3045 resulted in signicantly greater hip extensor impulse and signicantly greater
hamstrings and gluteus maximum activity. The greater hamstrings
activity during the lunge with the excessive forward trunk tilt likely
results in an increase in hamstrings force (loading the PCL and
unloading the ACL) compared to performing the lunge with a more
erect trunk. This is supported by relatively low peak hamstrings
force (approximately 150 N) in the current study compared to peak
quadriceps force (approximately 650 N). Therefore, technique differences while performing lunging exercises may partially explain
cruciate loading differences among lunging studies. Nevertheless,
it is clear from the squat and lunge literature that peak ACL force
magnitudes are relatively low (<75 N) and primarily generated at
lower knee angles between 0 and 30.
Compared to ACL forces during closed chain weight bearing
exercises, peak PCL forces are higher. In the current study peak
PCL forces were approximately 750800 N and occurred at higher
knee angles. Moreover, peak PCL forces were 400500 N during
the one leg squat (Escamilla et al., 2009), 700800 N during the
wall squat (Escamilla et al., 2009), and 18001900 N during the
barbell squat and leg press (Escamilla et al., 1998). However, because maximum strength of the healthy PCL is approximately
4000 N (Race and Amis, 1994), which is approximately twice as
strong as the healthy ACL, PCL forces are relatively low during
lunging and squatting type exercises. However, lunging and squatting exercises should be used cautiously during the early stages of
PCL rehabilitation, especially when performing the forward lunge
at higher knee angles, where PCL forces were greatest. Although
the reconstructed PCL graft typically has equal or greater ultimate
strength compared to the healthy PCL, it is unclear how much PCL
loading becomes injurious to the healing graft site during PCL rehabilitation, so caution should be employed.
It is unknown how much loading can safely occur in the reconstructed ACL (graft type must also be considered), although some
ACL loading during rehabilitation is probably desirable (Fitzgerald,
1997). Although the ultimate strength of the healthy ACL is
approximately 2200 N (Woo et al., 1991), and the ultimate
strength of the reconstructed ACL has been estimated between
approximately 25004000 N (Brown et al., 1993; Schatzmann
et al., 1998), it is unclear how much ACL loading becomes injurious
to the graft healing site during ACL rehabilitation.
An unexpected nding was that cruciate loading was not significantly different throughout the knee exion range between lunging with a stride compared to lunging without a stride. The
primary difference between with and without a stride occurred
during the initial portion of the descent phase, otherwise with
and without stride techniques were similar to each other. From Table 1, signicant differences in PCL forces were approached between the forward lunge and side lunge at 0 (P = 0.076), 10
(P = 0.036), and 20 (P = 0.096) of the descent phase. Moreover, between 0 and 20 of the descent phase, mean quadriceps forces derived from the biomechanical model were approximately twice as
great with a stride compared to without a stride. Large quadriceps
forces generated at low knee angles may result in ACL loading. This
occurs because force from the patellar tendon via the quadriceps
attempts to pull the tibia anterior at knee angles between 0 and
60 (Herzog and Read, 1993), which is restrained primarily by
the ACL (Butler et al., 1980). This may explain why PCL loading

was over 100 N less with a stride compared without a stride between 0 and 20 of the descent phase.
One reason quadriceps forces were greater with a stride compared to without a stride during the initial portion of the descent
phase is that the peak ground reaction forces acting on the lead
foot were approximately 20% greater with a stride compared to
without a stride between 0 and 20 knee angles of the descent.
The ground reaction force vector acting on the lead foot produced
a knee exor torque on the lead leg throughout the lunge that was
opposed by the knee extensors. Just after lead foot contact during
the descent, when the knee was exed 020, peak ground reaction
forces acting on the lead foot were greater with a stride because
the center of mass of the body had more forward and downward
acceleration compared to without a stride. Therefore, with a stride
the lead foot had to push harder into the ground to slow down the
forward and downward accelerating center of mass of the body
and control the rate of lead knee exion, which was approximately
45/s for both with and without stride conditions.
The current study examined healthy individuals without cruciate ligament pathology, because cruciate ligament forces during
forward and side lunge variations are currently unknown using
the healthy population. Additional research is warranted on patients with cruciate ligament pathology or reconstruction to determine if cruciate ligament forces generated during forward and side
lunge variations are similar between healthy individuals and patients with cruciate ligament deciencies or after ACL or PCL reconstruction. Addition studies are also needed using other techniques
variations, such as using a lunge step length shorter than the step
lengths used in the current study, which may help determine what
optimal step length minimizes cruciate ligament loading.
There are some limitations to the current study. Firstly, muscle
and cruciate ligament forces were estimated from biomechanical
modeling techniques and not measured directly, because it is currently not possible to measure cruciate ligament forces in vivo
while performing forward and side lunge exercises in healthy subjects. However, the cruciate force magnitudes from the current
study are similar to the cruciate force magnitudes reported during
the wall squat and one leg squat exercises (Escamilla et al., 2009).
Another limitation is that the current study was limited to sagittal
plane motion only, and future studies are needed to investigate the
effects of transverse plane rotary motions and frontal plane valgus/
varus motions on cruciate ligament loading. Future studies are also
needed to examine the effects of performing lunging exercises in
individuals with cruciate ligament deciencies.
5. Summary
Lunge technique variations do affect cruciate ligament loading.
Both forward and side lunges may be appropriate during ACL rehabilitation due to minimal ACL loading, especially the forward lunge.
However, clinicians should be cautious in prescribing the forward
and side lunge exercises during the early phases of PCL rehabilitation
when the graft site is still healing due to relatively high PCL forces,
especially at higher knee angles and during the forward lunge. PCL
forces were greater in the forward lunge compared to the side lunge
between 40 and 80 knee angles. No ACL loading was found during
both forward and side lunging. There were no differences in cruciate
ligament loading between with a stride and without a stride for all
knee angles. For each knee angles PCL forces were greater during
the ascent phase compared to the descent phase.
Acknowledgements
The efforts of Dr. Bonnie Raingruber and funding from the National Institute of Child Health and Human Developments Extra-

R.F. Escamilla et al. / Clinical Biomechanics 25 (2010) 213221

mural Associates Research Development Award program made this

research possible. Also acknowledged are Lisa Bonacci, Toni Burnham, Juliann Busch, Kristen DAnna, Pete Eliopoulos, and Ryan
Mowbray for all their assistance during data collection and
analyses.
Appendix A
Muscle force Fm(i) was calculated at each knee angle using the
following equation:
Fm(i) = ciklikviAirm(i)[EMGi/MVICi]  kv was calculated using the
following equations from Zajac (1989) and Epstein and Herzog
(1998):

kv b  a=F 0 v =b v concentric
kv C  C  1b a=F 0 v =b  v eccentric
with F0 representing isometric muscle force, l0 = muscle ber length
at rest; v = velocity, a = 0.32 F0, b = 3.2 l0/sec, and C = 1.8.
PCSA data from Wickiewicz et al. (1983) were used to determine the ratios of PCSA between muscle groups (Zheng et al.,
1998). According to Narici et al. (1988), the total PCSA (Ai) of the
quadriceps was approximately 160 cm2 for a 75 kg man. Total
PCSA of the quadriceps was scaled up or down by individual body
mass (Zheng et al., 1998). Forces generated by the knee exors and
extensors at MVIC were assumed to be linearly proportional to
their physiological cross-sectional area (Zheng et al., 1998). Muscle
force per unit physiological cross-sectional area at MVIC was
35 Ncm2 for the knee exors and 40 Ncm2 for the quadriceps
(Cholewicki et al., 1995; Narici et al., 1992, 1988; Wickiewicz
et al., 1984).
The objective function used to determine each ith muscles coefcient ci was as follows:
nm
nm
X
X
minf ci
1  ci 2 k T res 
T mi
i1

!2
;

i1

subject to clow 6 ci 6 chigh, where clow and chigh were lower and
upper limits for ci, and k was a constant. The weight factor c adjusted the nal muscle force calculation. The bounds on c were
set between 0.5 and 1.5, which resulted in the torques predicted
by the EMG driven model matching well (less than 2% of the knee
torques) with the torques generated from the inverse dynamics.
References
Balady, G., Berra, K., Golding, L., 2000. ACSMs guidelines for exercise testing and
prescription. Lippincott Williams & Wilkins, Baltimore.
Basmajian, J.V., Blumenstein, R., 1980. Electrode Placement in EMG Biofeedback.
Williams and Wilkins, Baltimore.
Beynnon, B.D., Johnson, R.J., Fleming, B.C., Stankewich, C.J., Renstrom, P.A., Nichols,
C.E., 1997. The strain behavior of the anterior cruciate ligament during
squatting and active exion-extension. A comparison of an open and a closed
kinetic chain exercise. Am. J. Sports Med. 25, 823829.
Brown Jr., C.H., Steiner, M.E., Carson, E.W., 1993. The use of hamstring tendons for
anterior cruciate ligament reconstruction. Technique and results. Clin. Sports.
Med. 12, 723756.
Butler, D.L., Noyes, F.R., Grood, E.S., 1980. Ligamentous restraints to anterior
posterior drawer in the human knee. A biomechanical study. J. Bone Joint Surg.
Am. 62, 259270.
Bynum, E.B., Barrack, R.L., Alexander, A.H., 1995. Open versus closed chain kinetic
exercises after anterior cruciate ligament reconstruction: a prospective
randomized study. Am. J. Sports Med. 23, 401406.
Cholewicki, J., McGill, S.M., Norman, R.W., 1995. Comparison of muscle forces and
joint load from an optimization and EMG assisted lumbar spine model: towards
development of a hybrid approach. J. Biomech. 28, 321331.

221

Epstein, M., Herzog, W., 1998. Theoretical Models of Skeletal Muscle: Biological and
Mathematical Considerations. John Wiley & Sons, New York.
Escamilla, R.F., Fleisig, G.S., Zheng, N., Barrentine, S.W., Wilk, K.E., Andrews, J.R.,
1998. Biomechanics of the knee during closed kinetic chain and open kinetic
chain exercises. Med. Sci. Sports Exerc. 30, 556569.
Escamilla, R.F., Fleisig, G.S., Zheng, N., Lander, J.E., Barrentine, S.W., Andrews, J.R.,
Bergemann, B.W., Moorman 3rd, C.T., 2001. Effects of technique variations on
knee biomechanics during the squat and leg press. Med. Sci. Sports Exerc. 33,
15521566.
Escamilla, R.F., Zheng, N., Imamura, R., Macleod, T.D., Edwards, W.B., Hreljac, A.,
Fleisig, G.S., Wilk, K.E., Moorman 3rd, C.T., Andrews, J.R., 2009. Cruciate
ligament force during the wall squat and one leg squat. Med. Sci. Sports
Exerc. 41, 408417.
Farrokhi, S., Pollard, C.D., Souza, R.B., Chen, Y.J., Reischl, S., Powers, C.M., 2008. Trunk
position inuences the kinematics, kinetics, and muscle activity of the lead
lower extremity during the forward lunge exercise. J. Orthop. Sports Phys. Ther.
38, 403409.
Fitzgerald, G.K., 1997. Open versus closed kinetic chain exercise: issues in
rehabilitation after anterior cruciate ligament reconstructive surgery. Phys.
Ther. 77, 17471754.
Ford, K.R., Myer, G.D., Hewett, T.E., 2003. Valgus knee motion during landing in high
school female and male basketball players. Med. Sci. Sports Exerc. 35, 1745
1750.
Heijne, A., Fleming, B.C., Renstrom, P.A., Peura, G.D., Beynnon, B.D., Werner, S., 2004.
Strain on the anterior cruciate ligament during closed kinetic chain exercises.
Med. Sci. Sports Exerc. 36, 935941.
Herzog, W., Read, L.J., 1993. Lines of action and moment arms of the major forcecarrying structures crossing the human knee joint. J. Anat. 182, 213230.
More, R.C., Karras, B.T., Neiman, R., Fritschy, D., Woo, S.L., Daniel, D.M., 1993.
Hamstrings an anterior cruciate ligament protagonist. An in vitro study. Am. J.
Sports Med. 21, 231237.
Morrison, J.B., 1970. The mechanics of the knee joint in relation to normal walking.
J. Biomech. 3, 5161.
Narici, M.V., Roi, G.S., Landoni, L., 1988. Force of knee extensor and exor muscles
and cross-sectional area determined by nuclear magnetic resonance imaging.
Eur. J. Appl. Physiol. 57, 3944.
Narici, M.V., Landoni, L., Minetti, A.E., 1992. Assessment of human knee extensor
muscles stress from in vivo physiological cross-sectional area and strength
measurements. Eur. J. Appl. Physiol. 65, 438444.
Ohkoshi, Y., Yasuda, K., Kaneda, K., Wada, T., Yamanaka, M., 1991. Biomechanical
analysis of rehabilitation in the standing position. Am. J. Sports. Med. 19, 605
611.
Pum, M.A., Shelburne, K.B., Torry, M.R., Decker, M.J., Pandy, M.G., 2004. Model
prediction of anterior cruciate ligament force during drop-landings. Med. Sci.
Sports Exerc. 36, 19491958.
Race, A., Amis, A.A., 1994. The mechanical properties of the two bundles of the
human posterior cruciate ligament. J. Biomech. 27, 1324.
Schatzmann, L., Brunner, P., Staubli, H.U., 1998. Effect of cyclic preconditioning on
the tensile properties of human quadriceps tendons and patellar ligaments.
Knee Surg. Sports Traumatol. Arthrosc. 6 (Suppl. 1), S56S61.
Shelburne, K.B., Pandy, M.G., 2002. A dynamic model of the knee and lower limb for
simulating rising movements. Comput. Methods Biomech. Biomed. Eng. 5, 149
159.
Shelburne, K.B., Pandy, M.G., Anderson, F.C., Torry, M.R., 2004. Pattern of anterior
cruciate ligament force in normal walking. J. Biomech. 37, 797805.
Shelburne, K.B., Torry, M.R., Pandy, M.G., 2005. Muscle, ligament, and joint-contact
forces at the knee during walking. Med. Sci. Sports Exerc. 37, 19481956.
Stuart, M.J., Meglan, D.A., Lutz, G.E., Growney, E.S., An, K.N., 1996. Comparison of
intersegmental tibiofemoral joint forces and muscle activity during various
closed kinetic chain exercises. Am. J. Sports Med. 24, 792799.
Toutoungi, D.E., Lu, T.W., Leardini, A., Catani, F., OConnor, J.J., 2000. Cruciate
ligament forces in the human knee during rehabilitation exercises. Clin.
Biomech. 15, 176187.
Wickiewicz, T.L., Roy, R.R., Powell, P.L., Edgerton, V.R., 1983. Muscle architecture of
the human lower limb. Clin. Orthop., 275283.
Wickiewicz, T.L., Roy, R.R., Powell, P.L., Perrine, J.J., Edgerton, V.R., 1984. Muscle
architecture and force-velocity relationships in humans. J. Appl. Physiol. 57,
435443.
Woo, S.L., Hollis, J.M., Adams, D.J., Lyon, R.M., Takai, S., 1991. Tensile properties of
the human femur-anterior cruciate ligament-tibia complex. The effects of
specimen age and orientation. Am. J. Sports Med. 19, 217225.
Yack, H.J., Collins, C.E., Whieldon, T.J., 1993. Comparison of closed and open kinetic
chain exercise in the anterior cruciate ligament-decient knee. Am. J. Sports
Med. 21, 4954.
Zajac, F.E., 1989. Muscle and tendon: properties, models, scaling, and application to
biomechanics and motor control. Crit. Rev. Biomed. Eng. 17, 359411.
Zheng, N., Fleisig, G.S., Escamilla, R.F., Barrentine, S.W., 1998. An analytical model of
the knee for estimation of internal forces during exercise. J. Biomech. 31, 963
967.