Chinese Journal of Traumatology 2011; 14(2):79-83 . 79 .

Stress changes of lateral collateral ligament at different

knee flexion with or without displaced movements: a 3-
dimensional finite element analysis
ZHONG Yan-lin, WANG You*, WANG Hai-peng, RONG Ke and XIE Le

【Abstract】 Objective: To create a 3-dimensional motions of anterior-posterior tibial translation, tibial varus
finite element model of knee ligaments and to analyse the rotation and internal-external tibial rotation were measured.
stress changes of lateral collateral ligament (LCL) with or Results: The maximum stress reached to 87%-113%
without displaced movements at different knee flexion versus natural stress in varus motion at early 30° of knee
conditions. flexions. The stress values were smaller than the peak value
Methods: A four-major-ligament contained knee speci- of natural stress at 0° (knee full extension) when knee bend-
men from an adult died of skull injury was prepared for CT ing was over 60° of flexion in anterior-posterior tibial trans-
scanning with the detectable ligament insertion footprints, lation and internal-external rotation.
locations and orientations precisely marked in advance. The Conclusion: LCL is vulnerable to varus motion in
CT scanning images were converted to a 3-dimensional almost all knee bending positions and susceptible to ante-
model of the knee with the 3-dimensional reconstruction rior-posterior tibial translation or internal-external rotation
technique and transformed into finite element model by the at early 30° of knee flexions.
software of ANSYS. The model was validated using experi- Key words: Knee joint; Collateral ligaments; Finite
mental and numerical results obtained by other scientists. element analysis
The natural stress changes of LCL at five different knee
Chin J Traumatol 2011; 14(2):79-83
flexion angles (0°, 30°, 60°, 90°, 120°) and under various

B
eing one of the biggest and most complicated an in vitro knee model which could simulate the knee
joints of human body, knee joint plays a key motion for the studies of knee ligament kinematics and
role in the motion of lower extremity and is biomechanics.
subjected to a high incidence of injuries. For the assess-
ment of the functional behaviour of knee ligaments and With various hypotheses of human knee kinematics,
the intimate understanding of their injury mechanisms, many knee models1-3 have been designed for different
the biomechanics of the human knee has been investi- applications. And with the development of technology,
gated extensively in the past. Considering the com- a more accurate and precise model is request by sci-
plexity of in vivo study, most of the researchers utilized entists and surgeons. The 3-dimensional (3-D) finite el-
ement technique is one of the most important advances
in biomechanic study of the knee. It combines the com-
DOI: 10.3760/cma.j.issn.1008-1275.2011.02.003
puter 3-D reconstruction technology with the elasticity
Department of Orthopedic Surgery, Shanghai 9th
People’s Hospital, School of Medicine, Shanghai Jiaotong
theory. The researched subject or target is considered
University, Shanghai 200011, China (Zhong YL, Wang Y, as a consecutive elastic whole and divided into a se-
Wang HP and Rong K) ries of limited physical units called finite elements. With
National Die and Mold CAD Engineering Research this method, we can set up a special 3-D finite element
Center, Shanghai Jiaotong University, Shanghai 200011, knee model loaded with the interesting targets (as knee
China (Xie L)
ligaments) to investigate their mechanical property more
*Corresponding author: Tel: 86-21-63139920. Email:
drwangyou@163.com
accurately under different given conditions.
This study was supported by the grant from the Na-
tional High Technology Research and Development Pro- Recently, there have been lots of 3-D finite element
gram of China (863 Program, No.2006AA01Z310). studies on the knee ligaments of the anterior cruciate liga-
. 80 . Chinese Journal of Traumatology 2011; 14(2):79-83

ment (ACL)4 and medial collateral ligament (MCL)5, gram for conducting 3-D reconstruction of the geomet-
showing great advantages in repeatability, accuracy and ric patterns of the knee-partial ligament models (without
resemblance to the human physiological environment. ligament insertions) at the different knee flexion angles.
In regard to the lateral collateral ligament (LCL), The femoral or tibial 3-D skeleton model with clear foot-
however, the reports are few in the literature. The lack prints of four main ligament insertions was matched
of LCL data may be partially due to its low rupture with the bony part of each 3-D knee-partial ligament
occurrence. In an epidemiological study of total 500 model by generating close-fits of driven screw sites in
knee injuries, the rate of LCL injury was only 4%, com- both models, facilitating the entire ligament reconstruc-
pared with 63% for ACL injury and 44% for MCL injury.6 tion (with insertions) and the completion of 3-D knee-
Nevertheless, LCL injury especially complete rupture ligament models (Figure 1).
injury, is far more difficult to heal clinically and often
requires surgical intervention. A thorough investigation Finite element models
of the dynamic function of LCL in knee stability during The geometric patterns of the 3-D knee-ligament
full range of motion is still necessary for understanding models were transferred into finite element models by
its injury mechanism and proper ways of prevention. In setting the property of unit, generating the knot and
this article, a 3-D finite element model of knee with its dividing meshes. For simplification, the bony parts of
four main ligaments derived from the CT scanning image the models (femur and tibia) were recognized as rigid-
was established and the stress of LCL was gained by ity and the ligaments were described as successive,
finite element analysis during passive knee bending un- uniform, isotropic and liner elastic in property. The
der the different conditions of knee movement positions. Young’s modulus of (345.0±22.4) MPa was used for
LCL, ACL, PCL 7 and (332.2±58.3) MPa for MCL8 as a
METHODS material property and loaded to the finite element mod-
els according to the literature. In our model, we as-
Preparation of the specimen sumed that there was no difference in the material be-
A normal knee of a fresh corpse died of skull injury tween the ligament and its insertion, and the material
was used for study. All the periarticular soft tissues characteristics depending on time, such as viscoelas-
including skin and muscle were removed from the knee ticity and relaxation were neglected. The ligament in-
joint. The distal femur, proximal tibia and the four major sertion sites, considered as the contact surface, were
ligaments, i.e. ACL, posterior cruciate ligament (PCL), glued with the ligaments by the software. After setting
MCL and LCL, were kept intact. Three 2.8 mm screws up the property of unit and the contact surface, the 3-D
were positioned at the bone surfaces of the femur and knee-ligament finite element model was formed auto-
tibia respectively as markers for the possible matching matically by dividing grid software at 5 different flexion
in later study. The ligaments were carefully wreathed angles. The finite element model contains the femur,
by steel wires 1 mm in diameter and 5 mm in interval. tibia and four major ligaments of knee joint (Figure 2).
The steel wires were used to locate the positions and In order to compare our results to those of other
shapes of ligaments between their insertions on CT scientists’, a 134 N anterior load was applied to the
scanning images. tibia at the full extension, and the tibial movement was
recorded by the software.
CT scanning and 3-D reconstruction
CT scanning at intervals of 0.625 mm was performed Stress measurement and analysis
for 5 times on the ligament-contained knee specimen In 5 models without knee displacement, the abso-
at the flexion angle of 0°, 30°, 60°, 90° and 120° lute lengths of LCL at different knee flexion angles (0°,
respectively. The bony specimen underwent CT scan- 30°, 60°, 90° and 120°) were measured first. The center
ning once again after removal of all the ligaments and of the insertion site was determined using the ANSYS
marked in the edges of the ligament insertion footprints software. Cut the ligaments into a few cross-sections
by multiple puncture with 1.5 mm K-wire on the femur and divide their center separately through the software
and tibia for the 3-D skeleton models. The CT scanning CATIA. Connect the center of the insertion site and the
data of ligament-contained knee specimens were then cross-sections by the curve which can represent the 3-
introduced into the Mimics 10.01 medical imaging pro- D length of the ligament (Figure 3). The model with the
Chinese Journal of Traumatology 2011; 14(2):79-83 . 81 .

shortest length of LCL was considered as reference different knee displacement motions, i.e. 5 mm ante-
model. Then we compared the lengths of other 4 mod- rior tibial translation, 5 mm posterior tibial translation,
els with those of the reference model and calculated 5° tibial varus rotation, 5° tibial internal rotation and 5°
their relative stress values of LCL by a finite element tibial external rotation. These movements were taken
software, setting the stress value of the reference model at 5 knee flexion angles (0°, 30°, 60°, 90° and 120°)
as 0. A series of stress values of LCL were also studied and their stress values in the ligament were recorded.
using the same reference model when we simulate 5

Figure 1. The 3-D model of knee joint at different flexion angles.

knee flexion respectively. And a 5.19 mm anterior move-

ment was obtained under the 134 N anterior tibial load.

Stress of LCL with various knee displacements

Varus rotation of 5°of the knee The results of
the LCL stress under varus rotation of the knee are shown
in Figure 4. After 5° varus rotation, the LCL stress rose
dramatically. The highest stress was 8.31 MPa at 0° of
the knee flexion, more than twice of the value of the
initial peak stress in the non-displaced model. The added
value of the stress compared with initial value of the same
flexion angles was 4.0-4.4 MPa. Though the changing
tendency was similar to that of the initial curve which
Figure 2. The finite element model including the femur, tibia, and showed a decrease with the knee flexion, the stress of
four major ligaments of knee joint. Figure 3. The conceptual dia-
the ligament was 3.95 MPa at 120° of the knee flexion,
gram of the calculate method to the knee ligaments.
being still much higher than the initial peak stress at 0°.

RESULTS
Anterior or posterior tibial translation of 5 mm
After 5 mm anterior or posterior tibial translation, the
Stress of LCL without knee displacement
stress on the LCL increased moderately (Figures 5 and 6).
The length of LCL was found to decrease with the
The highest stress appeared at the models of 0° knee
knee getting more flexion and the shortest length of
flexion, being 6.17 MPa and 5.92 MPa respectively.
LCL among 5 non-displaced models was 47.63 mm in
With the increase of flexion angle of the knee, the added
the model of 120° knee flexion. Supposing the model of
value of stress in either anterior or posterior displace-
120° knee flexion was the reference model and the
ment models dropped gradually from about 2 MPa in
stress of LCL was 0, the stress of LCL was estimated
knee fully extended models to less than 1 MPa in 120°
to be 1.13 MPa in the model of 90° knee flexion, 1.87
knee flexion models. The stress values were smaller
MPa in the model of 60° knee flexion, 2.95 MPa in the
than the peak value of initial stress at 0° when knee
model of 30° knee flexion and 3.90 MPa in the model of 0°
bending was over 60° of flexion.
. 82 . Chinese Journal of Traumatology 2011; 14(2):79-83

Internal or external rotation of 5° of the knee Fig- DISCUSSION

ures 7 and 8 show the results of stress changes at
different knee flexion with 5° internal or external rotation In this article, we presented a finite element model
of the knee. Stress on the LCL increased moderately of knee with four major ligaments. In order to verify the
and was estimated to be 5.02 MPa and 5.30 MPa at 0° model and compare the results with those of other
models. As the knees flexed more, the increased stress authors’, a 134 N anterior-posterior load was applied to
of internal or external rotation models dropped rapidly the tibia at full extension. Under the anterior tibial load
to 0.40 MPa and 0.59 MPa respectively from the initial of 134 N at full extension, the results of the other au-
levels at 120°. Both stress values were much lower than thors were as follows: Gabriel et al 9 got a 5.0 mm
the initial peak stress of non-displaced model at 0° when displacement; Fu et al10, 5.2 mm; Song et al11, 4.6 mm;
knee flexion was over 60°. and Pena et al12, 4.75 mm. Our result about the ante-
rior tibial displacement was 5.19 mm under the condi-
tion of 134 N load, which was similar to the data ob-
tained by other authors. Thereupon our finite model can
be proved to be valid, and then we can use this model
to simulate the knee motions.

Figure 4. The stress on the LCL at different flexion angles with

LCL is a main supporting structure outside of the
tibial varus rotation compared with the initial stress.
knee. It extends from the lateral epicondyle to the top
outside surface of the fibular head. As we know, LCL is
taut in knee extension and relaxed in knee flexion.
During knee flexion in neutral rotation the distance be-
tween the femoral and fibular attachment sites of the
lateral collateral ligament was found to decrease to 88%
Figure 5. The stress on the LCL at different flexion angles with of its value in full extension.6 Our study revealed that
anterior tibial translation compared with the initial stress. the stress of LCL decreased by about 19.9%-29.0%
with knee having each 30° bending from full extension
to 120° flexion. From Figure 2 to Figure 6, we can find
that no matter which movement was carried out to dis-
place the knee, the increased stress on LCL peaked in
full extension and dropped gradually with the knee bend-
ing toward 120°. Therefore, we can infer that LCL brings
Figure 6. The stress on the LCL at different flexion angles with
its functions into full play as knee extension. Any ex-
posterior tibial translation compared with the initial stress.
ternal moment of force such as pure varus, anterior
translation, posterior translation, internal rotation or
external rotation, which further stretches LCL as knee
full extension, would damage the ligament.

Our study also demonstrated the difference of LCL

Figure 7. The stress on the LCL at different flexion angles with stress increase under different conditions of knee move-
tibial internal rotation compared with the initial stress. ment positions. As compared with the initial peak stress
of non-displaced model, the varus motion created great
stress increase on LCL throughout the knee range of
motion, while anterior/ posterior translation, or internal/
external rotation could only result in moderate or slight
stress increase at 0° and 30° of knee flexions. The
maximum increase rates reached to 87%-113% in varus
Figure 8. The stress on the LCL at different flexion angles with motion at early 30° of knee flexions. These data were
tibial external rotation compared with the initial stress. 58% or 52% in anterior or posterior translation motions
Chinese Journal of Traumatology 2011; 14(2):79-83 . 83 .

and 29% or 31% in internal or external rotation motions REFERENCES

at 0° of knee flexion. The result is consistent with the
previous findings that LCL serves as the primary stabi- 1. Toutoungi DE, Lu TW, Leardini A, et al. Cruciate ligament
lizer against varus force and the important stabilizer forces in the human knee during rehabilitation exercises. Clin
against rotation force. The decreased role in its stability Biomech 2000;15(3):176-187.
function in deep flexion of the knee would be compen- 2. Zheng N, Fleisig GS, Escamilla RF, et al. An analytical
sated by other structures in lateral side of the knee. model of the knee for estimation of internal forces during exercise.
J Biomech 1998;31(10):963-967.
In fact, isolated LCL injuries are rare. More commonly, 3. Li G, Suggs J, Gill T. The effect of anterior cruciate ligament
this ligament is injured combined by PLC injury or ACL/ injury on knee joint function under a simulated muscle load: a
PCL injury.13 It has been reported by other authors that three-dimensional computational simulation. Ann Biomed Eng
LCL can resist the posterior displacement.6,12 The rise 2002;30(5):713-720.
in LCL stress at full extension of the knee in our study 4. Hirokawa S, Tsuruno R. Three-dimensional deformation
when an anterior or posterior tibial translation was ap- and stress distribution in an analytical/computational model of the
plied could partially explain some clinical findings of anterior cruciate ligament. J Biomech 2000;33(9):1069-1077.
combined ACL/PCL injuries of LCL without obvious varus 5. Gardiner JC, Weiss JA. Subject-specific finite element analy-
angulation. Regarding knee twisting injury mechanism, sis of the human medial collateral ligament during valgus knee
PLC was proven to be the main structure for resisting loading. J Orthop Res 2003;21(6):1098-1106.
external rotation.6 As the secondary stabilizer, LCL helps 6. Meister BR, Michael SP, Moyer RA, et al. Anatomy and
PLC in preventing knee from external rotation, and there- kinematics of the lateral collateral ligament of the knee. Am J
fore they are commonly subjected to combined injuries. Sports Med 2000;28(6):869-878.
However, we found that the high stress appeared only 7. Butler DL, Kay MD, Stouffer DC. Comparison of material
at the full extension and decreased quickly with the properties in fascicle-bone units from human patellar tendon and
knee flexion. This finding supports the view that LCL knee ligaments. J Biomech 1986;19(6):425-432.
contributes to the external rotatory stability at early 8. Quapp KM, Weiss JA. Material characterization of human
range of knee flexion. The similar high stress of LCL in medial collateral ligament. J Biomech Eng 1998;120(6):757-763.
internal rotation indicates that it may also play some 9. Gabriel MT, Wong EK, Woo SL, et al. Distribution of in
role in assisting medial posterior corner to restrain knee situ forces in the anterior cruciate ligament in response to rotatory
internal rotation. loads. J Orthop Res 2004;22(1):85-89.
10. Fu FH, Bennett CH, Lattermann C, et al. Current trend in
This model contains main structures of the knee anterior cruciate ligament reconstruction. Part 1: biology and bio-
(femur, tibia, and four major ligaments), and can simu- mechanics of reconstruction. Am J Sports Med 1999;27(6):821-
late different knee movements through the software. The 830.
lack of muscle and capsule, together with neglecting 11. Song Y, Debski RE, Musahl V, et al. A three-dimensional
the restraint roles of menisci and cartilage, are the limi- finite element model of the human anterior cruciate ligament: a
tations of this study, which may lead to an overestima- computational analysis with experimental validation. J Biomech
tion of the ligament force. The actual isotropic, nonlin- 2004;37(3): 383-390.
ear and nonuniform structure properties of knee com- 12. Pena E, Calvo B, Martínez MA, et al. A three-dimensional
ponents also result in some discrepancy from in vitro finite element analysis of the combined behavior of ligaments and
situation due to various conditions and load settings. menisci in the healthy human knee joint. J Biomech 2006;39(9):
However, to our knowledge, this knee model is more 1686-1701.
accurate not only in anatomically reconstructing four 13. Paletta GA, Warren RF. Knee injuries and Alpine skiing.
major ligaments but also in mimicing kinematics of Treatment and rehabilitation. Sports Med 1994;17(6):411-423.
passive knee bending, compared with other models of
previous studies. The relationship between stress (Received May 12, 2010)
changes in LCL and various external applied motions Edited by SONG Shuang-ming
at different knee flexion angles described in the study
is therefore more reliable.