Clinical Biomechanics 24 (2009) 842–849

Contents lists available at ScienceDirect

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

Biomechanical analysis of posterior cruciate ligament retaining high-flexion

total knee arthroplasty
J. Zelle a,*, A.C. Van der Zanden a, M. De Waal Malefijt a, N. Verdonschot a,b
a
Orthopaedic Research Laboratory, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands
b
Laboratory for Biomechanical Engineering, University of Twente, Enschede, The Netherlands

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

Article history: Background: High-flexion knee replacements have been developed to accommodate a large range of
Received 6 April 2009 flexion (>120°) after total knee arthroplasty. Both posterior cruciate ligament retaining and sacrificing
Accepted 9 August 2009 high-flexion knee designs have been marketed. The main objective of this study was to evaluate the bio-
mechanical performance of a cruciate-retaining high-flexion knee replacement. Furthermore, the
mechanical behaviour of this high-flexion knee replacement was compared to both a cruciate-retaining
Keywords: conventional and a posterior-stabilized high-flexion knee replacement.
Total knee arthroplasty
Methods: A finite element prosthetic knee model was developed to analyze the mechanical performance
Finite element analysis
High-flexion
of the knee designs evaluated in this study. Polyethylene stresses and the amount of femoral rollback
Posterior-stabilized were studied during a squatting movement (flexion 6 150°).
Cruciate-retaining Findings: During deep knee flexion, the cruciate-retaining high-flexion design demonstrated a lower peak
tibio-femoral contact stress (74.7 MPa) than the cruciate-retaining conventional design (96.5 MPa). The
posterior-stabilized high-flexion design showed the lowest peak tibio-femoral contact stress at the con-
dylar articulation (54.2 MPa), although the post was loaded higher (77.4 MPa). The knee designs analyzed
in this study produced a similar amount of femoral rollback during normal knee flexion (flexion 6 120°),
whereas the cruciate-retaining designs showed a paradoxical anterior movement of the femoral condyles
during high-flexion (flexion > 120°).
Interpretation: The current study demonstrates a cruciate-retaining high-flexion knee replacement pro-
duces a lower prosthetic load than a conventional cruciate-retaining replacement during deep knee flex-
ion. Compared to a posterior-stabilized high-flexion design, the cruciate-retaining high-flexion design
demonstrated an equivalent prosthetic loading along with an inferior amount of femoral rollback in
the high-flexion range. Posterior cruciate ligament balancing is an important surgical aim for high-flexion
knee arthroplasty.
Ó 2009 Elsevier Ltd. All rights reserved.

1. Introduction flexion angles such as squatting, kneeling and sitting cross-legged

(Mulholland and Wyss, 2001). Recently, high-flexion knee implants
The traditional goals of total knee arthroplasty (TKA) are pain have been developed to provide for a larger RoM (flexion > 120°)
relief and restoration of normal knee function. In addition, post- after TKA than conventional implants.
operative range of motion (RoM) has gained interest over the last High-flexion knee implants are often based on successful con-
decade. Several clinical studies have demonstrated that standard ventional designs. In most cases, the posterior sagittal femoral
TKA components allow knee flexion up to 120° of flexion (Banks geometry has been adapted to prevent the occurrence of poly-
et al., 2003) which is not adequate for deep knee flexion activities ethylene edge loading during deep knee flexion. Examples of
such as squatting and kneeling. Hence, active knee patients experi- high-flexion knee replacements are the Nexgen Flex (Zimmer,
ence limitations during these activities (Weiss et al., 2002). Fur- Warsaw, IN, USA), the Scorpio Flex (Stryker, Kalamazoo, MI,
thermore, the TKA market is flourishing in non-Western USA), the Genesis II High-Flex (Smith and Nephew, Memphis,
countries, where daily living activities frequently involve large TN, USA) and the PFC Sigma RP-F (Depuy, Warsaw, IN, USA).
Apart from differences in design regarding the flexion range,
knee implants also vary in the ability to substitute for posterior
* Corresponding author. Address: Orthopaedic Research Laboratory, Radboud cruciate ligament (PCL) deficiency. Therefore, orthopaedic com-
University Nijmegen Medical Centre, P.O. Box 9101, 6500 HB Nijmegen, The
Netherlands.
E-mail address: J.Zelle@orthop.umcn.nl (J. Zelle).

0268-0033/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.clinbiomech.2009.08.004
 J. Zelle et al. / Clinical Biomechanics 24 (2009) 842–849 843

panies developed both PCL-retaining and PCL-sacrificing high- 2.1. Finite element knee model
flexion knee replacements to meet the surgical demands. In
PCL-sacrificing TKA designs, also referred to as posterior- 2.1.1. Model description
stabilized designs, a post-cam feature replaces the function of The three-dimensional dynamic FE knee model used in this
the PCL. study has previously been described (Barink et al., 2008; Zelle
The PCL, conserved during cruciate-retaining TKA, has been et al., 2009) and runs using the FE program Marc (MSC Software
described to prevent posterior translation of the tibia and to Corporation, Santa Ana, CA, USA). The FE knee model consisted of
contribute to femoral rollback during knee flexion (Williams a proximal tibia and fibula, TKA components, a quadriceps and pa-
and Thornhill, 2005). Femoral rollback or posterior translation tella tendon, collateral ligaments, a non-resurfaced patella and a
of the femoral condyles is essential to maximize knee flexion PCL (Fig. 1).
and to avoid posterior bone-implant impingement during deep In total three rotating platform TKA designs (Fig. 2) from the
knee flexion (Banks et al., 2003). In earlier studies, it has been same manufacturer (Depuy, Warsaw, IN, USA) were mechanically
demonstrated that posterior-stabilized designs lead to a more evaluated using the FE knee model. Firstly, the TKA components
substantial and predictable amount of femoral rollback than of a newly designed cruciate-retaining high-flexion knee replace-
cruciate-retaining designs (Argenson et al., 2005). Furthermore, ment were integrated in the knee model (PFC Sigma CR150). Sec-
physiological PCL strain patterns in cruciate-retaining TKA ondly, both a cruciate-retaining conventional (PFC Sigma CR) and
have proven to be difficult to reproduce after surgery (Scuderi a posterior-stabilized high-flexion knee replacement (PFC Sigma
and Pagnano, 2001). The PCL should not be too tight or too RP-F) were included in the FE knee model. The TKA systems were
lax after surgery in order to obtain an adequate amount of all incorporated following the surgical procedure using 0° of tibial
femoral rollback (Arima et al., 1998). A tight PCL may also slope.
lead to detrimental posterior polyethylene wear (Swany and Hexahedral (brick) elements were used to model the osseous
Scott, 1993). Obtaining physiological PCL functioning is a chal- tissues (tibia, fibula, patella) and the TKA components. However,
lenging surgical aim especially after cruciate-retaining high- the femoral component was modelled as a rigid body using only
flexion TKA, requiring more femoral rollback than conventional the surface description. The geometries of osseous tissues were
TKA. taken from a CT-scan and the geometries of the prosthetic com-
In the past, kinematic studies compared cruciate-retaining ponents were taken from CAD files provided by the manufac-
with posterior-stabilized conventional knee replacements (Haas turer. Since it was outside the scope of this study to analyze
et al., 2002; Victor et al., 2005). Recently, kinematic studies also bone-implant interaction, a simplified distribution of bone prop-
analyzed cruciate-retaining high-flexion TKA systems in knee erties was included in the model. Homogeneous linear material
patients (Cates et al., 2008; Suggs et al., 2009) and by using properties of the cortical (E = 19 GPa) and trabecular bone
cadaveric measurements (Most et al., 2005). In addition, finite (E = 120 MPa) were obtained from bone mineral density infor-
element (FE) models have proven their use to analyze TKA load- mation derived from the CT-scan and the tibial tray was mod-
ing in the normal flexion range (RoM 6 120°; Godest et al., 2002; elled as a linear material (E = 210 GPa). The polyethylene was
Halloran et al., 2005) and the high-flexion range (RoM > 120°; modelled as an elastic–plastic non-linear material having an
Morra and Greenwald, 2005; D’Lima et al., 2008). However, no elastic modulus of 974 MPa and an initial Von Mises yield stress
FE models have been reported that focus on the effect of PCL of 14.0 MPa. Nearly ideal plasticity was assumed for the poly-
functioning on the prosthetic knee load and kinematics during ethylene beyond Von Mises stress values of 29.5 MPa (Kurtz
high-flexion. et al., 1998). Knee ligaments and tendons were modelled as
The objective of this study was to investigate the effect of PCL composite materials consisting of two-dimensional shell ele-
retention on the biomechanical performance of a high-flexion knee ments and non-linear line elements to represent the solid ma-
replacement during deep knee flexion (RoM 6 150°). Since FE mod- trix and collagen fibres of the ligaments, respectively. The
els are adequate to analyze the prosthetic knee loading under linear shell elements (E = 10 MPa) were used in the contact def-
weight-bearing conditions, an FE knee model was developed to inition to allow for ligament-wrapping around bony and poly-
analyze the mechanics of a cruciate-retaining high-flexion knee ethylene surfaces. Collagen fibres of the different ligaments
replacement. Both the amount of femoral rollback and the polyeth- and tendons (excluding the PCL) were modelled as non-linear
ylene stresses were evaluated for this implant type. Furthermore, materials having a quadratic ‘toe-region’ and linear material
the mechanical performance of the cruciate-retaining high-flexion behaviour above a certain linear strain level according to Blank-
knee replacement was compared to the behaviour of both a cruci- evoort et al. (1991b). The elastic modulus of the patella tendon
ate-retaining conventional and a posterior-stabilized high-flexion (643.1 MPa), the lateral and medial collateral ligament (345.0
TKA design. More specifically, the main research questions and 332.1 MPa) and the linear strain level (e  0.06) were de-
addressed in this study were whether a cruciate-retaining rived from experimental data (Butler et al., 1986; Quapp and
high-flexion TKA design outperforms a cruciate-retaining conven- Weiss, 1998).
tional TKA design during deep knee flexion (RoM > 120°) and Frictionless tibio-femoral and patello-femoral contact were de-
whether a cruciate-retaining high-flexion TKA design demon- fined in the FE knee model as well as contact between the tibial
strates a similar mechanical performance as a posterior-stabilized tray and the polyethylene. A sensitivity study showed that friction
high-flexion design in terms of polyethylene stresses and femoral had little effect on the polyethylene stress distributions. Further-
rollback. more, the tibial tray was fixed on the tibia using glued contact,
mimicking cemented tibial fixation.

2. Methods 2.1.2. Boundary conditions

The boundary conditions applied to the knee model were: (1)
An FE model of the prosthetic knee joint was used to analyze fixation of the femoral component, (2) fixation of the proximal
TKA kinematics and kinetics during deep knee flexion. The FE knee ends of the quadriceps tendon and (3) application of the ground
model has been used in earlier studies to analyze posterior-stabi- reaction force (Fig. 1). During each simulation the tibia was loaded
lized knee replacements only. For use in this study, the PCL was at the ankle level using a ground reaction force of 350 N (subject
modelled and implemented in the FE knee model. mass  70 kg) and knee flexion was achieved by releasing the
844 J. Zelle et al. / Clinical Biomechanics 24 (2009) 842–849

Fig. 1. The three-dimensional dynamic FE model used in this study. Different parts of the knee model have been pointed out as well as the boundary conditions (fixation of
the quadriceps tendon, fixation of the femoral component and application of the ground reaction force Fgrf).

Fig. 2. The three knee replacements analyzed in this study: PFC Sigma CR (a), PFC Sigma CR150 (b) and PFC Sigma RP-F (c). Design differences in posterior condylar geometry
are clearly visible.

quadriceps tendon incrementally. Hence, the knee model was free 2.1.3. Convergence analysis
to determine its own kinematics during a simulated squatting The effect of mesh density changes on the calculated polyethyl-
movement (RoM 6 150°; flexion velocity  12°/s). Tibial rotation ene stress distributions was evaluated using a simplified model of
was limited by a torsion spring (j = 0.37 Nm/deg), replicating fric- the high-flexion PFC Sigma RP-F in extension (Fig. 3a). This implant
tion between the foot and the ground. Inclusion of this rotational was considered to be representative for the three TKA designs ana-
restraint resulted in maximal 5–10° internal rotation during flex- lyzed in this study since the sagittal geometry of the different de-
ion reasonably similar to clinical observations. The model was signs was very similar in the normal flexion range. In this
numerically damped (c = 0.01 Ns/m) to prevent vibrations and convergence analysis, the femoral component was compressed
numerical instabilities from occurring during the simulations. Trial on the polyethylene using an axial force of 3600 N, identical to ear-
simulations showed this numerical damping had virtually no ad- lier polyethylene contact pressure measurements analyzing the
verse effects on the prosthetic load and kinematics. Each increment same TKA design (Shiramizu et al., 2009). Both the peak Von Mises
of knee flexion was considered to be converged when a conver- stress and the peak contact stress were evaluated for several poly-
gence ratio of 0.01 was reached (residual forces 6 0.01  external ethylene mesh densities (element edge length = 1, 2, 2.5, 5 and
forces). 10 mm). The peak Von Mises stress and contact stress converged
 J. Zelle et al. / Clinical Biomechanics 24 (2009) 842–849 845

Fig. 3. Simplified TKA model used in the convergence analysis (a) and the peak contact and Von Mises stress for several polyethylene mesh densities (b). The peak Von Mises
stress smoothly converged to 20 MPa for smaller mesh densities, whereas the peak contact stress was somewhat more sensitive to local mesh discontinuities.

to roughly 20 MPa (Fig. 3b), which also has been reported by Shir- Experimental data from both tests were used to fit the SLS model
amizu et al. (2009). Small edge lengths (61 mm) were found to be parameters (E1, E2, g1) representing an average PCL. An extra dam-
more sensitive to local geometrical discontinuities slightly increas- per had to be added to the SLS model (g2) to better match the
ing the peak contact stress. Hence, an edge length in the range of experimental data at higher strain rates. In order to implement fi-
2–2.5 mm was used to model the TKA inserts. bre recruitment, the SLS model parameters were made depending
linearly on the strain level (Eq. (2)).
 
2.2. Posterior cruciate ligament model E1 de g1 dr de
r ¼ E1 e þ g1 1 þ  þ g2 ð1Þ
E2 dt E2 dt dt
2.2.1. Material model
The PCL is a complex structure having non-linear material prop- E1 ¼ k1 e þ k2 ¼ 129:0e þ 1:3
erties. In the current study, the most important PCL requirements E2 ¼ k3 e ¼ 24:0e
were a non-linear stress–strain behaviour due to fibre recruitment ð2Þ
g1 ¼ k4 e ¼ 1659:0e
and strain rate dependency (viscoelasticity). It should be noted
that velocities and accelerations were relatively low in the FE knee
g2 ¼ k5 e ¼ 24:0e
model.
Spring-damper systems are frequently used in constitutive 2.2.2. Anatomy
models to implement viscoelastic material behaviour. Examples The PCL modelled in this study was divided into an antero-lat-
of spring-damper systems are the Maxwell model, Kelvin–Voigt eral and a postero-medial bundle. Similar to the other knee liga-
model and the Standard Linear Solid model (Van der Vegt, 2002). ments, the PCL bundles consisted of line elements to implement
For use in this study, the Standard Linear Solid (SLS) was deemed non-linear stress–strain behaviour and shell elements used in the
appropriate since it is the simplest model to describe viscoelastic contact definition. Femoral and tibial attachment sites of the PCL
behaviour including creep, stress relaxation and recovery and has bundles and bundle thickness were obtained from anatomical
been used by other research groups to model soft tissues measurements (Takahashi et al., 2006).
(Machiraju et al., 2006). The SLS model is shown (Fig. 4) together
with its constitutive relationship (Eq. (1)).
Pioletti et al. (1998) performed both tension tests at multiple 2.2.3. PCL zero force length
strain rates and stress relaxation tests using human PCL specimens. PCL balancing is another important factor to account for in the
FE knee model used in this study. Variations in PCL laxity likely oc-
cur after surgery and determine the flexion angle at which the PCL
starts to conduct force. The PCL should not be too tight or too lax in
the FE knee model, which obviously affects the polyethylene stress
level and the amount of femoral rollback. Mommersteeg et al.
(1995) studied the smallest ligament length at which the PCL starts
to generate force during tension tests: the PCL zero force length
(L0). Based on this earlier study and some trial simulations, three
different PCL zero force lengths (30.4, 34.0 and 37.6 mm) were
implemented in the FE knee model to study the effect of PCL laxity
on the TKA load during normal knee flexion (RoM 6 120°). Inclu-
sion of the PCL zero force length in the FE knee model redefined
the PCL strain used in the SLS model by evaluating whether the
PCL length was larger or smaller than the prescribed PCL zero force
length (Eq. (3)).
LðtÞ < L0 ! eðtÞ ¼ 0
Fig. 4. Standard Linear Solid material model used to implement the viscoelastic LðtÞ  L0 ð3Þ
behaviour of the PCL. Extra damping was added to better match experimental data LðtÞ  L0 ! eðtÞ ¼
at higher strain rates. L0
846 J. Zelle et al. / Clinical Biomechanics 24 (2009) 842–849

2.3. Finite element analysis outcome parameters cruciate-retaining designs the peak Von Mises stress was located
at the condylar contact interface of the insert. However, for the
Based on the analysis of PCL laxity, a PCL having a zero force posterior-stabilized high-flexion design the highest stress peaks
length of 34.0 mm was considered being balanced (see Section were found at the post-cam articulation. Furthermore, the total
3.1) and was used in the biomechanical comparison of the three Von Mises stress distribution area at final knee flexion was much
TKA designs. The knee replacements were evaluated during a sim- larger for the posterior-stabilized high-flexion design due to the
ulated squatting movement (45° 6 RoM 6 150°) and three out- additional weight-bearing contact area (post-cam interaction).
come parameters were assessed concerning the polyethylene
loading: (1) the peak Von Mises stress, (2) the peak tibio-femoral
3.3. Peak contact stress
contact stress and (3) the amount of femoral rollback. Both peak
Von Mises and contact stress could be derived directly from the
In the normal flexion range (flexion 6 120°) the three TKA de-
FE stress distributions. Femoral rollback was evaluated by studying
signs demonstrated equivalent contact stress patterns as expected
the antero-posterior (AP) position of the centres of pressure on the
due to the nearly identical sagittal geometry in this flexion range
medial and lateral compartment of the insert with respect to the
(Fig. 7a). The peak contact stress was approximately 40 MPa for
posterior edge of the polyethylene. Subsequently, the AP-positions
all TKA designs at 120° of flexion. During high-flexion
of the medial and lateral centres of pressure were averaged to ob-
(flexion > 120°), the conventional PFC Sigma CR demonstrated the
tain one single positional value.
highest contact stress at the condylar contact interface
(96.5 MPa), followed by the PFC Sigma CR150 (74.7 MPa) and the
3. Results PFC Sigma RP-F (54.2 MPa). However, the peak contact stress at
the post observed for the PFC Sigma RP-F (77.4 MPa) was very sim-
3.1. PCL zero force length ilar to stress values encountered by the cruciate-retaining designs
(Fig. 7a). The posterior femoral condyles of the PFC Sigma CR
The various PCL zero force lengths included in the FE knee mod- started to dig into the polyethylene during high-flexion, increasing
el caused the PCL to start conducting force at different flexion an- the polyethylene contact stress and level of deformation.
gles (flexion  50°, 70°, 90°; Fig. 5a). The highest maximal PCL
force (1755 N) was observed for the shortest zero force length 3.4. Femoral rollback
(30.4 mm) and considerably lower maximal PCL forces (859 and
402 N) were observed for the other zero force lengths of In general, the three TKA systems all demonstrated femoral roll-
34.0 mm and 37.6 mm, respectively (Fig. 5a). In addition, the max- back in the normal flexion range (Fig. 7b). The AP-position of the
imal PCL force affected the peak polyethylene contact stresses, tibio-femoral contact points was comparable for the three TKA sys-
which were 56.3 MPa, 35.6 MPa and 38.1 MPa for the different zero tems at 120° of flexion: approximately 10 mm from the posterior
force lengths in increasing order. Simulations with the tightest PCL edge of the insert. However, the posterior-stabilized knee design
even led to the observation of ‘booking’ of the polyethylene during demonstrated more anterior movement in the beginning of the
which the plastic insert lifted off anteriorly from the tibial tray flexion cycle than the cruciate-retaining knee implants. Post-cam
(Fig. 5b). interaction clearly initiated at 100° of flexion. During deep knee
flexion (flexion > 120°), both cruciate-retaining TKA designs dem-
3.2. Peak Von Mises stress onstrated paradoxical anterior movement of the tibio-femoral con-
tact points, neutralizing the amount of rollback produced in the
Both trend and magnitude of the peak Von Mises stress were normal flexion range. The posterior-stabilized high-flexion design
relatively similar for all TKA designs due to plastic deformation lev- continued to roll backwards during deep knee flexion (final AP-po-
elling off the Von Mises stress. Overall, the peak Von Mises stress sition = 7.6 mm). The anterior movement of the femoral condyles
increased with the flexion angle and reached peak values of 29– during deep knee flexion was somewhat smaller for the conven-
30 MPa at the maximal knee flexion angle (Fig. 6). In case of the tional cruciate-retaining design compared to the high-flexion

Fig. 5. Differences in PCL force due to variations in zero force length (a) and ‘booking’ of the polyethylene due to a tight PCL (b).
 J. Zelle et al. / Clinical Biomechanics 24 (2009) 842–849 847

Fig. 6. Von Mises stress distribution plots (a–c) and AP cross-sections of the lateral polyethylene compartment (d–f) at 150° of flexion. From left to right: PFC Sigma CR (a and
d), PFC Sigma CR150 (b and e) and PFC Sigma RP-F (c and f).

Fig. 7. The peak contact stress (a) and the antero-posterior contact point location (b) for the three TKA designs analyzed in this study.

cruciate-retaining design due to the fact that the femoral condyles and posterior-stabilized designs. The cruciate-retaining high-flex-
dug into the polyethylene. ion TKA design produced the highest polyethylene stresses at the
condylar contact interface, whereas the posterior-stabilized high-
flexion design generated the highest stresses at the post. Although
4. Discussion the TKA designs analyzed in this study all demonstrated femoral
rollback in the normal flexion range (RoM 6 120°), we found the
The aim of the current study was to evaluate the effect of PCL femoral condyles of the cruciate-retaining TKA designs to move
retention on the biomechanical performance of a high-flexion knee anterior during deep knee flexion. This femoral roll forward can
replacement. For this purpose, a PCL was modelled and included in be considered as a potential drawback for cruciate-retaining
an FE knee model containing a cruciate-retaining high-flexion knee high-flexion TKA and it therefore seems highly important to bal-
replacement. Furthermore, the mechanical performance of the cru- ance the PCL with care.
ciate-retaining high-flexion knee replacement was compared to The FE knee model set-up as used in this study has been utilized
the behaviour of both a cruciate-retaining conventional and a pos- in similar cadaveric knee tests, often based on the Oxford knee
terior-stabilized high-flexion TKA design. testing rig (Zavatsky, 1997). Although the FE knee model can be
In summary, the polyethylene stress level of the cruciate- considered representative for the in vivo situation, the knee model
retaining high-flexion design during deep knee flexion (RoM > contains several limitations which possibly affected the outcome of
120°) was reduced with respect to the cruciate-retaining conven- this study. Firstly, the ground reaction force was assumed to act
tional design and was comparable to that produced by the poster- through the ankle joint, while the ground reaction force may be lo-
ior-stabilized high-flexion design. These results indicate that PCL cated more anterior in reality during heels-down squatting
retention does not necessarily lead to an increased loading of the (Thambyah, 2008). Hence, the knee flexion moment generated by
polyethylene insert during deep knee flexion compared to poster- our FE knee model may be somewhat higher compared to the
ior-stabilized TKA. However, the location where the highest poly- in vivo situation. However, the polyethylene contact stress calcu-
ethylene stresses occurred differed between the PCL-retaining lated in this study was comparable to other studies (Morra and
848 J. Zelle et al. / Clinical Biomechanics 24 (2009) 842–849

Greenwald, 2005; D’Lima et al., 2008; Shiramizu et al., 2009). Sec- It therefore seems recommendable to pay extra attention to intra-
ondly, only the quadriceps forces were assumed to balance the operative PCL balancing for patients that are anticipated to use
flexion moment of the squatting posture. This obviously is a sim- their implant at high flexion angles.
plification of reality, although Dahlkvist et al. (1982) showed that
this may be a reasonable approximation. Thirdly, in an earlier
study we evaluated the effect of thigh-calf contact during deep 5. Conclusions
knee flexion and we demonstrated that this posterior soft-tissue
compression has a considerable effect on the TKA load (Zelle The current study demonstrates that a cruciate-retaining high-
et al., 2009). In this study, thigh-calf contact was not included in flexion knee replacement produces a lower prosthetic load than a
the FE knee model. Therefore, the polyethylene loading during conventional cruciate-retaining replacement during deep knee
high-flexion was somewhat overestimated by the FE knee model. flexion (RoM > 120°). Compared to a posterior-stabilized high-flex-
Overall, the FE knee model with all its strengths and limitations ion design, the cruciate-retaining high-flexion design demon-
was used only for comparative analyses and mainly qualitative ef- strated an equivalent prosthetic loading. However, the amount of
fects were analyzed in this study. femoral rollback produced by the cruciate-retaining high-flexion
Balancing the PCL is a challenging surgical aim during TKA and, design was inferior compared to the posterior-stabilized high-flex-
as this study showed, can have important implications if not done ion design. Posterior cruciate ligament balancing is an important
correctly. The effect of variations in PCL laxity as occurring after surgical aim for high-flexion TKA.
TKA was evaluated by inclusion of several PCL zero force lengths
in the FE knee model according to Mommersteeg et al. (1995). Acknowledgements
Based on the PCL laxity analysis, the outcome generated with a
PCL zero force length of 34.0 mm appeared best-congruent with This research was made possible by a research Grant of Depuy
the literature. This specific zero force length caused the PCL to start International Ltd., Leeds, UK.
to function at approximately 70° of flexion and resulted in a max-
imal PCL force of 859 N. Earlier studies reported that the PCL has a
strength in the range of 739–1878 N (Race and Amis, 1994) and References
that the flexion angle at which the PCL starts to conduct force Argenson, J.N., Scuderi, G.R., Komistek, R.D., Scott, W.N., Kelly, M.A., Aubaniac, J.M.,
ranges from 60° to 70° of flexion (Blankevoort et al., 1991a; Mar- 2005. In vivo kinematic evaluation and design considerations related to high
kolf et al., 2006). Hence, the PCL having a zero force length of flexion in total knee arthroplasty. J. Biomech. 38, 277–284.
Arima, J., Whiteside, L.A., Martin, J.W., Miura, H., White, S.E., McCarthy, D.S., 1998.
34.0 mm was considered to be an adequately balanced PCL in our
Effect of partial release of the posterior cruciate ligament in total knee
FE knee model. However, one should realize that this particular arthroplasty. Clin. Orthop. Relat. Res. 353, 194–202.
PCL zero force length may not be directly usable in other knee Banks, S., Bellemans, J., Nozaki, H., Whiteside, L.A., Harman, M., Hodge, W.A., 2003.
models as its appropriate length heavily depends on the knee anat- Knee motions during maximum flexion in fixed and mobile-bearing
arthroplasties. Clin. Orthop. Relat. Res. 410, 131–138.
omy and TKA components considered. Barink, M., De Waal Malefijt, M., Celada, P., Vena, P., Van Kampen, A., Verdonschot,
Inclusion of an excessively tight PCL led to increased polyethyl- N., 2008. A mechanical comparison of high-flexion and conventional total knee
ene stresses and the observation of ‘booking’ of the plastic insert, arthroplasty. Proc. Inst. Mech. Eng. [H] 222, 297–307.
Blankevoort, L., Huiskes, R., De Lange, A., 1991a. Recruitment of knee joint
which has been described in clinical studies (Shannon et al., ligaments. J. Biomech. Eng. 113, 94–103.
2007). The extent to which an inappropriately balanced PCL is able Blankevoort, L., Kuiper, J.H., Huiskes, R., Grootenboer, H.J., 1991b. Articular contact
to remodel shortly after TKA is uncertain. Some authors report that in a three-dimensional model of the knee. J. Biomech. 24, 1019–1031.
Butler, D.L., Kay, M.D., Stouffer, D.C., 1986. Comparison of material properties in
excessive pre-stress would quickly diminish due to viscoelastic fascicle-bone units from human patellar tendon and knee ligaments. J. Biomech.
properties and biological adaptation (Markolf et al., 1996). In any 19, 425–432.
case, it seems advisable to optimally balance the PCL in the in- Callaghan, J.J., Insall, J.N., Greenwald, A.S., Dennis, D.A., Komistek, R.D., Murray,
D.W., Bourne, R.B., Rorabeck, C.H., Dorr, L.D., 2001. Mobile-bearing knee
tra-operative situation, which may be facilitated using balancing replacement: concepts and results. Instr. Course Lect. 50, 431–449.
instruments (Christen et al., 2007) or computer navigation Cates, H.E., Komistek, R.D., Mahfouz, M.R., Schmidt, M.A., Anderle, M., 2008. In vivo
techniques. comparison of knee kinematics for subjects having either a posterior stabilized
or cruciate retaining high-flexion total knee arthroplasty. J. Arthroplast. 23,
The paradoxical anterior movement of the femoral condyles as
1057–1067.
demonstrated in this study has been reported in clinical studies as Christen, B., Heesterbeek, P., Wymenga, A., Wehrli, U., 2007. Posterior cruciate
well (Dennis et al., 1998). Terms such as ‘paradoxical anterior ligament balancing in total knee replacement: the quantitative relationship
movement’ or ‘roll forward’ have frequently been used in the liter- between tightness of the flexion gap and tibial translation. J. Bone Joint Surg. Br.
89, 1046–1050.
ature (Callaghan et al., 2001; Scuderi, 2005). Conversely, there are Dahlkvist, N.J., Mayo, P., Seedhom, B.B., 1982. Forces during squatting and rising
also studies which did report the occurrence of femoral rollback from a deep squat. Eng. Med. 11, 69–76.
during deep knee flexion (Most et al., 2005; Cates et al., 2008). Dennis, D.A., Komistek, R.D., Colwell Jr., C.W., Ranawat, C.S., Scott, R.D., Thornhill,
T.S., Lapp, M.A., 1998. In vivo anteroposterior femorotibial translation of total
However, one should be careful with comparing kinematic studies knee arthroplasty: a multicenter analysis. Clin. Orthop. Relat. Res. 356, 47–57.
to one another as the definition of femoral rollback differs between D’Lima, D.D., Steklov, N., Fregly, B.J., Banks, S.A., Colwell Jr., C.W., 2008. In vivo
studies. Some papers determine the amount of femoral rollback by contact stresses during activities of daily living after knee arthroplasty. J.
Orthop. Res. 26, 1549–1555.
analyzing the movement of the tibio-femoral contact points (Den- Godest, A.C., Beaugonin, M., Haug, E., Taylor, M., Gregson, P.J., 2002. Simulation of a
nis et al., 1998; Banks et al., 2003), while others analyze the move- knee joint replacement during a gait cycle using explicit finite element analysis.
ment of the epicondylar axis during knee flexion (Most et al., J. Biomech. 35, 267–275.
Haas, B.D., Komistek, R.D., Stiehl, J.B., Anderson, D.T., Northcut, E.J., 2002. Kinematic
2005). In the simulations with the cruciate-retaining designs we comparison of posterior cruciate sacrifice versus substitution in a mobile
clearly found the femoral condyles sliding forward during deep bearing total knee arthroplasty. J. Arthroplast. 17, 685–692.
knee flexion, whereas the epicondylar axis remained more or less Halloran, J.P., Petrella, A.J., Rullkoetter, P.J., 2005. Explicit finite element modeling of
total knee replacement mechanics. J. Biomech. 38, 323–331.
in place. The fact that we did not see a considerable amount of roll-
Kurtz, S.M., Ochoa, J.A., White, C.V., Srivastav, S., Cournoyer, J., 1998. Backside
back in the cruciate-retaining high-flexion design case may jeopar- nonconformity and locking restraints affect liner/shell load transfer
dize the utility of this type of knee design at high-flexion mechanisms and relative motion in modular acetabular components for total
conditions. Current knowledge shows that rollback is of eminent hip replacement. J. Biomech. 31, 431–437.
Machiraju, C., Phan, A.V., Pearsall, A.W., Madanagopal, S., 2006. Viscoelastic studies
importance to provide enough space to allow for deep knee flexion of human subscapularis tendon: relaxation test and a Wiechert model. Comput.
without posterior bone-implant impingement (Banks et al., 2003). Methods Programs Biomed. 83, 29–33.
 J. Zelle et al. / Clinical Biomechanics 24 (2009) 842–849 849

Markolf, K.L., Burchfield, D.M., Shapiro, M.M., Davis, B.R., Finerman, G.A., Shannon, F.J., Cronin, J.J., Cleary, M.S., Eustace, S.J., O’Byrne, J.M., 2007. The posterior
Slauterbeck, J.L., 1996. Biomechanical consequences of replacement cruciate ligament-preserving total knee replacement: do we ‘preserve’ it? A
of the anterior cruciate ligament with a patellar ligament allograft. Part I: radiological study. J. Bone Joint Surg. Br. 89, 766–771.
insertion of the graft and anterior–posterior testing. J. Bone Joint Surg. Am. 78, Shiramizu, K., Vizesi, F., Bruce, W., Herrmann, S., Walsh, W.R., 2009. Tibiofemoral
1720–1727. contact areas and pressures in six high flexion knees. Int. Orthop. 33, 403–406.
Markolf, K.L., Feeley, B.T., Jackson, S.R., McAllister, D.R., 2006. Biomechanical studies Suggs, J.F., Kwon, Y.M., Durbhakula, S.M., Hanson, G.R., Li, G., 2009. In vivo flexion
of double-bundle posterior cruciate ligament reconstructions. J. Bone Joint Surg. and kinematics of the knee after TKA: comparison of a conventional and a high
Am. 88, 1788–1794. flexion cruciate-retaining TKA design. Knee. Surg. Sports Traumatol. Arthrosc.
Mommersteeg, T.J., Blankevoort, L., Huiskes, R., Kooloos, J.G., Kauer, J.M., Hendriks, 17, 150–156.
J.C., 1995. The effect of variable relative insertion orientation of human knee Swany, M.R., Scott, R.D., 1993. Posterior polyethylene wear in posterior cruciate
bone–ligament-bone complexes on the tensile stiffness. J. Biomech. 28, 745– ligament-retaining total knee arthroplasty. A case study. J. Arthroplast. 8, 439–
752. 446.
Morra, E.A., Greenwald, A.S., 2005. Polymer insert stress in total knee designs during Takahashi, M., Matsubara, T., Doi, M., Suzuki, D., Nagano, A., 2006. Anatomical study
high-flexion activities: a finite element study. J. Bone Joint Surg. Am. 87 (Suppl. of the femoral and tibial insertions of the anterolateral and posteromedial
2), 120–124. bundles of human posterior cruciate ligament. Knee Surg. Sports Traumatol.
Most, E., Li, G., Sultan, P.G., Park, S.E., Rubash, H.E., 2005. Kinematic analysis of Arthrosc. 14, 1055–1059.
conventional and high-flexion cruciate-retaining total knee arthroplasties: an Thambyah, A., 2008. How critical are the tibiofemoral joint reaction forces during
in vitro investigation. J. Arthroplast. 20, 529–535. frequent squatting in Asian populations? Knee 15, 286–294.
Mulholland, S.J., Wyss, U.P., 2001. Activities of daily living in non-Western cultures: Van der Vegt, A.K., 2002. From Polymers to Plastics, first ed. VSSD, Delft.
range of motion requirements for hip and knee joint implants. Int. J. Rehabil. Victor, J., Banks, S., Bellemans, J., 2005. Kinematics of posterior cruciate ligament-
Res. 24, 191–198. retaining and -substituting total knee arthroplasty: a prospective randomised
Pioletti, D.P., Rakotomanana, L.R., Benvenuti, J.F., Leyvraz, P.F., 1998. Viscoelastic outcome study. J. Bone Joint Surg. Br. 87, 646–655.
constitutive law in large deformations: application to human knee ligaments Weiss, J.M., Noble, P.C., Conditt, M.A., Kohl, H.W., Roberts, S., Cook, K.F., Gordon, M.J.,
and tendons. J. Biomech. 31, 753–757. Mathis, K.B., 2002. What functional activities are important to patients with
Quapp, K.M., Weiss, J.A., 1998. Material characterization of human medial collateral knee replacements? Clin. Orthop. Relat. Res. 404, 172–188.
ligament. J. Biomech. Eng. 120, 757–763. Williams, T.J., Thornhill, T.S., 2005. The technique of PCL retention in total knee
Race, A., Amis, A.A., 1994. The mechanical properties of the two bundles of the arthroplasty. In: Bellemans, J., Ries, M.D., Victor, J. (Eds.), Total Knee
human posterior cruciate ligament. J. Biomech. 27, 13–24. Arthroplasty: A Guide to Get Better Performance. pp. 177–181.
Scuderi, G.R., 2005. The stiff total knee arthroplasty: causality and solution. J. Zavatsky, A.B., 1997. A kinematic-freedom analysis of a flexed-knee-stance testing
Arthroplast. 20, 23–26. rig. J. Biomech. 30, 277–280.
Scuderi, G.R., Pagnano, M.W., 2001. Review article: the rationale for posterior Zelle, J., Barink, M., De Waal, M.M., Verdonschot, N., 2009. Thigh-calf contact: does it
cruciate substituting total knee arthroplasty. J. Orthop. Surg. (Hong Kong) 9, 81– affect the loading of the knee in the high-flexion range? J. Biomech. 42, 587–
88. 593.