You are on page 1of 6

Jason P.

Halloran
Department of Mechanical and Materials Engineering, Computational Biomechanics Laboratory, University of Denver, 2390 South York, Denver, CO 80208

Chadd W. Clary Lorin P. Maletsky


Department of Mechanical Engineering, University of Kansas, 1530 W 15th Street, Lawrence, KS 66045

Verication of Predicted Knee Replacement Kinematics During Simulated Gait in the Kansas Knee Simulator
Evaluating total knee replacement kinematics and contact pressure distributions is an important element of preclinical assessment of implant designs. Although physical testing is essential in the evaluation process, validated computational models can augment these experiments and efciently evaluate perturbations of the design or surgical variables. The objective of the present study was to perform an initial kinematic verication of a dynamic nite element model of the Kansas knee simulator by comparing predicted tibioand patellofemoral kinematics with experimental measurements during force-controlled gait simulation. A current semiconstrained, cruciate-retaining, xed-bearing implant mounted in aluminum xtures was utilized. An explicit nite element model of the simulator was developed from measured physical properties of the machine, and loading conditions were created from the measured experimental feedback data. The explicit nite element model allows both rigid body and fully deformable solutions to be chosen based on the application of interest. Six degrees-of-freedom kinematics were compared for both tibio- and patellofemoral joints during gait loading, with an average root mean square (rms) translational error of 1.1 mm and rotational rms error of 1.3 deg. Model sensitivity to interface friction and damping present in the experimental joints was also evaluated and served as a secondary goal of this paper. Modifying the metal-polyethylene coefcient of friction from 0.1 to 0.01 varied the patellar exion-extension and tibiofemoral anterior-posterior predictions by 7 deg and 2 mm, respectively, while other kinematic outputs were largely insensitive. DOI: 10.1115/1.4001678 Keywords: TKR, nite element model, kinematics, knee mechanics, contact mechanics

Mark Taylor
Bioengineering Sciences Research Group, University of Southampton, Southampton SO17 1BJ, UK

Anthony J. Petrella
DePuy, a Johnson & Johnson Company, 700 Orthopaedic Drive, Warsaw, IN 46581

Paul J. Rullkoetter1
Department of Mechanical and Materials Engineering, Computational Biomechanics Laboratory, University of Denver, 2390 South York, Denver, CO 80208 e-mail: prullkoe@du.edu

Introduction

Characterizing kinematics and contact stresses in total knee replacements can lend insight into potential clinical performance. Experimental knee simulators have been developed to quantify the total knee replacement kinematics as well as the resulting polyethylene wear. Physical knee wear simulators have generally included multiple stations with tibial and femoral replacement components 1,2 . In addition, single station hip- and ankle-loaded test frames have been developed and typically include both tibio- and patellofemoral articulations 36 . Although these simulators are used for evaluation of implant mechanics, parametric evaluation of the implant geometry, alignment, or loading conditions is generally too time consuming and expensive to t within the design phase of development. Investigating variations in the implant geometry is especially costly due to the requirement of producing physical prototypes. Given the time constraints, implant designers have integrated validated computational models as a viable alternative because of the ease with which geometry, relative alignment, or loading variables can be modied. Many models of the total knee replacement
1 Corresponding author. Contributed by the Bioengineering Division of ASME for publication in the JOURNAL OF BIOMECHANICAL ENGINEERING. Manuscript received September 25, 2006; nal manuscript received April 6, 2010; accepted manuscript posted April 28, 2010; published online July 1, 2010. Assoc. Editor: Avinash Patwardhan.

articulation have been formulated using nite element or rigid body dynamics methods 711 . Recently, explicit nite element models of the total knee replacement that simultaneously predict joint kinematics and contact mechanics have been developed and veried by direct comparison with kinematics data from forcecontrolled joint simulators 12,13 . In these explicit nite element analyses, tibiofemoral kinematics, and contact pressure and area were predicted during gait loading, and anterior-posterior A-P translation and internal-external I-E rotation were in good agreement with experimental measurements. The explicit nite element models are unique in that a fully deformable or rigid body representation of the implant may be used. It was previously shown that for gait loading, rigid body analyses predicted relative kinematics and contact pressures that were good estimates of fully deformable analyses, and required only a fraction of the time approximately 6 min compared with 68 h 13 . However, if internal stresses/strains are desired, a fully deformable representation must be analyzed. In addition to the tibiofemoral modeling, verication of isolated patellofemoral kinematics was performed using experimental data from the hipand ankle-loaded Purdue knee simulator 13 . In these prior models 13 , each articulation was developed and veried independently, and thus, there was no verication of the predicted kinematics during simultaneous tibio- and patellofemoral loadings. As the tibio- and patellofemoral mechanics are interdependent 1419 , simultaneous kinematic prediction is critical AUGUST 2010, Vol. 132 / 081010-1

Journal of Biomechanical Engineering

Copyright 2010 by ASME

Downloaded 27 Jul 2011 to 147.46.230.42. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

Fig. 1 Experimental Kansas knee simulator left and nite element representation with implant mesh right. Applied loads and exion angle are represented by arrows. Boundary conditions include the axial load, hip exion angle, quadriceps load, internal-external torque, and adduction-abduction force.

for accurate evaluation of the total knee replacement components. Thus, the goal of the present study was to perform an initial comparison of the predicted and experimental tibio- and patellofemoral kinematics during simulated gait loading using a dynamic nite element model of the hip- and ankle-loaded Kansas knee simulator KKS . For this study, the total knee replacement components were mounted in aluminum xtures to ensure a known, carefully controlled loading environment. The xtured analysis was chosen to isolate the constraint provided by the implant and to evaluate the nite element model representation without the uncertainty introduced by testing implanted cadavers. The secondary objective of this study was to evaluate the sensitivity of the model predictions to the metal-polyethylene coefcient of friction and representation of damping in the physical joints of the experimental conguration. This study serves as an important step in the validation process of developing predictive computational tools for the evaluation of implant performance.

Fig. 2 Applied loading and exion angle as a function of the gait cycle. 060% gait cycle represents stance phase 0% equals heal strike and 60100% gait cycle represents the swing phase. The same is true for all gures with percent gait cycle on the abscissa.

Methods

2.1 Experimental Kinematic Analysis. Experimental kinematic analysis was performed using the KKS with a current xedbearing, semiconstrained posterior cruciate-retaining PCR total knee replacement PFC Sigma size 3, DePuy, a Johnson & Johnson company, Warsaw, IN during gait simulation. The experimental system is briey described here; more detailed information may be found in Ref. 6 . The KKS is a servohydraulic system that uses actuators to impart a vertical force at the hip, and a medial-lateral force and tibial torque about a vertical axis at the ankle Fig. 1 . The ankle complex on the simulator is a universal joint mounted on a linear sled. This mechanical arrangement allows the ankle joint to ex and extend, adduct or abduct, and rotate internally or externally, while the linear sled can translate in the medial-lateral direction and allow varus-valgus rotation at the knee. The hip joint allows exion-extension and inferiorsuperior motions, and is constrained in all other degrees of freedom. A quadriceps actuator mounted on the femur balances the applied vertical load through the patellar tendon in the experiment, a Kevlar band . The quadriceps actuator was operated in displacement control to duplicate a specied hip exion angle prole Fig. 2 , and the required quadriceps force was measured at the tibial tuberosity attachment. Physical dimensions, masses, mass moments of inertia, and centers of mass were measured from the simulator. Distance from the ankle complex to the joint line is approximately 400 mm and from the joint line to the hip is 470 mm. The quadriceps actuator is aligned to recreate an ap081010-2 / Vol. 132, AUGUST 2010

propriate q-angle 7 deg . For this testing, actuator loading proles were utilized to recreate level gait 20 . Force and displacement feedback data for each actuator was recorded throughout the testing to obtain accurate input proles for the corresponding nite element models. In total, 10 experimental cycles were run and averaged to obtain the input proles for the nite element model, with the rst and last cycles not included. Loading between cycles was very consistent with variations for each experimental cycle within 1% of the average. Fixtured brackets were machined to mount the implant. The tibial component was aligned with 0 deg posterior tilt, perpendicular to the long axis, and 0 deg internal-external and varus-valgus rotations as referenced from the at distal surface of the insert and axes of symmetry. For displacement degrees of freedom, the tibial component geometric center was placed in the center of the long axis of the aluminum xture representing the tibia. The femoral component was aligned with 0 deg external, exion, and varusvalgus rotations as referenced from the at interior bone cut surfaces , and neutrally aligned with the tibial component in the anterior-posterior and medial-lateral directions. The domed patellar component was mounted ush with the Kevlar strap used in the extensor mechanism and adjusted to lie in the center of the trochlear groove at the start of the experiments. Before each test, the articular surface of the implant was lubricated with petroleum jelly. Tibio- and patellofemoral kinematics were measured using an Optotrak system Northern Digital, Waterloo, Ontario, Canada . The Optotrak system utilizes infrared emitting diodes IRED to track the rigid body motion of the femur, patella, and tibia for the Transactions of the ASME

Downloaded 27 Jul 2011 to 147.46.230.42. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

test cycles. Transformation data between each rigid body were then transformed into clinically relevant coordinate frames and described using a three-cylindric model of the joint motion 21 . Output kinematics were averaged for the middle 8 cycles, with insignicant variations standard deviations of 0.09 deg and 0.14 mm realized between cycles. 2.2 Finite Element Model of the Kansas Knee Simulator. The nite element model of the KKS was developed in ABAQUS/Explicit Abaqus, Inc., Providence, RI from computeraided design CAD models of the implant and the measured physical properties of the simulator Fig. 1 . The measured lengths, masses, and inertial properties of the experimental simulator were included in the nite element model, and the experimental geometry not essential for the analysis was simplied. Translational and rotational joint element combinations were used to model the machine ankle and hip joints. As in the experiment, the hip sled was allowed to translate in the inferior-superior direction and rotate in exion-extension. All other femoral degrees of freedom were constrained. The ankle joint accommodates exionextension, adduction-abduction, and internal-external rotations, and is attached to a linear connector that allows medial-lateral translation. The extensor mechanism was modeled with a combination of connector elements. The quadriceps tendon and patellar ligament were attached to the anterior side of the patellar button and approximated the stiffness of the Kevlar band used. The quadriceps actuator was modeled with a connector element to impart a force along its axis and simulate the action of the hydraulic actuator used in the experiment. Although friction and damping are present in the complex experimental joints, these parameters are difcult to obtain, and the model was run initially without representation of these losses. Small medial-lateral oscillation was observed at the ankle joint in these initial kinematic analyses, and subsequently, a sensitivity study was performed to understand the sensitivity of the model to damping present at the medial-lateral sled. In order to nominally damp this degree of freedom, the coefcient was increased in coarse of 0.05 N s/mm increments until the oscillation subsided. Using the coarse grid results, a ner increment of 0.01 N s/mm was subsequently evaluated to obtain the nal nominal damping value. Three-dimensional, eight-noded brick nite elements were used to represent the polyethylene tibial insert and patellar button. Approximately 8000 elements were used to model the inserts, with 16,000 elements for the patellar button. Element edge lengths were approximately 1.3 mm for the insert and 0.7 mm for the patella button. Previous mesh density studies veried convergence for contact pressure and stress results 13 . The femoral component was modeled using approximately 21,000 three-dimensional, rigid triangular elements. Rigid body analysis with an optimized pressure-overclosure relationship was used in this study 22 , while deformable analyses were performed to verify rigid body kinematic predictions. For the deformable analyses, a nonlinear elastic-plastic material model was used for the elements representing the polyethylene tibial insert and patellar button 23 . Contact was modeled using a penalty-based method with a weight factor. As a result, contact forces are dened as a function of the penetration of the master surface into the slave surface. To evaluate the sensitivity of the model to friction at the metal-polyethylene interface, both tibio- and patellofemoral , the coefcient of friction was varied from 0.01 to 0.1. 2.3 Computational Kinematic Analysis. The nite element model was run using boundary conditions to replicate the experimental loading; measured machine feedback was used for all inputs to the nite element model, including the hip exion angle, axial load, tibial torque, adduction-abduction load, and quadriceps load. Hip exion-extension was controlled, and resulted in a range of knee exion of approximately 60 deg Fig. 3 . The axial load and hip exion angle were applied at the hip, while the adductionabduction load and torque were applied at the ankle. Journal of Biomechanical Engineering

Fig. 3 Model and experimental tibiofemoral exion +-extension F-E and varus +-valgus V-V rotations as a function of the gait cycle. The results represent a coefcient of friction of 0.08, and the experimental results represent 8 averaged cycles.

A three-cylindric model of knee motion was utilized to describe the six degrees-of-freedom DOF relative kinematics of the tibioand patellofemoral articulations 21 . Embedded coordinate frames in the patellar button, femur, and insert were represented using nodes, and their positions were monitored throughout the gait cycle. A post-processing script was used to interface with the output databases of the nite element models, implement the model of knee motion, and describe the relative motions as three rotations and three translations. Patellofemoral rotations include exion-extension, which was described about the tibiofemoral exion axis 24 , internal-external tilt, which occurs about the long axis of the patella, and patellar spin, which occurs about a oating axis perpendicular to both the exion and patellar long axes. Patellar translations were described simply as the inferiorsuperior, medial-lateral, and anterior-posterior motions of the patella in the femoral anatomical frame. In order to accurately compare the kinematics between the model and experiment, the embedded coordinate frames were placed in identical positions with respect to the components. Contact pressure and area were also monitored throughout the nite element simulations, but no experimental data were available for comparison. To allow for comparison of the experimental and model kinematics, the root mean square error was calculated for each degree of freedom.

Results

3.1 Tibiofemoral Kinematics. Model-predicted and experimental tibiofemoral exions achieved a rms error of less than 0.45 deg for either rigid or deformable representations Fig. 3, rigid analysis shown . As this degree of freedom is essentially controlled by the applied hip exion angle, this comparison is essentially a verication of the implementation of the anatomical coordinate frames and description of joint motion. Both experimental and predicted varus-valgus rotation of the knee resulting from the applied adduction-abduction force at the ankle were very small, with ranges of less than 2.1 deg each Fig. 3, rigid analysis shown . Predicted I-E tibiofemoral rotation with no damping in the model included a small oscillation between approximately 65% and 80% of the gait cycle when damping in the ankle joint was neglected . A subsequent sensitivity study demonstrated that the I-E prediction during this 15% of the cycle was sensitive to damping at the ankle, but that the other tibio- and patellofemoral kinematics were not substantially affected. Example results from an addition of nominal damping of 0.53 N s/mm at the ankle eliminated the oscillation during the swing phase and are shown AUGUST 2010, Vol. 132 / 081010-3

Downloaded 27 Jul 2011 to 147.46.230.42. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

Fig. 4 Experimental and predicted tibiofemoral internal +-external I-E rotation as a function of the gait cycle. Data include rigid body analyses with zero and nominal ankle dampings, as well as a deformable analysis with nominal ankle damping all results are with a coefcient of friction of 0.08.

Fig. 6 Model and experimental patellofemoral rotations as a function of the gait cycle + rotations are extension, internal tilt, and varus spin; all results are with a coefcient of friction of 0.08

Fig. 4 . The deformable analysis had a rms error of 1.5 deg when compared with the experiment and a range of 12.2 deg, while the experimental I-E range was 14.5 deg. Comparing the results between the rigid and deformable results shows differences primarily at the peak rotation peak rms error of up to 1.4 deg . The peak rotations of around 10 deg for all data sets represent an internal rotation of the tibia with respect to the femur. For the range tested, the coefcient of friction for metal-polyethylene interaction did not affect the predicted tibiofemoral rotations results substantially. Model results for the tibiofemoral A-P motion were within 2 mm of the experimental data throughout the simulations for either rigid or deformable representations Fig. 5 , and were somewhat sensitive to the coefcient of friction. The experimental data achieved a range of 10.4 mm and, as the friction coefcient increased, the model-predicted ranges increased from 11 mm to 13 mm. The rms error values comparing the predicted and experimental data ranged from 0.5 mm rigid body model using a coefcient of friction of 0.01 to 0.94 mm rigid body model using a

coefcient of friction of 0.1 . While the A-P motion was sensitive to the tibiofemoral friction coefcient, the maximum errors only increased from 1.6 mm to 3.2 mm for the range of friction evaluated. The single deformable analysis performed with a coefcient of friction of 0.08 had a rms error of less than 0.44 mm when comparing the deformable and rigid body data. The deformable analysis resulted in a range of 12 mm and a rms error, with respect to the experimental data, of 0.8 mm. Maximum A-P motion occurred at the deepest exion achieved by the model and experiment, and represents a posterior motion of the femur with respect to the tibia femoral rollback . The tibiofemoral A-P data should not be thought of as the position of relative tibiofemoral contact, but as tibiofemoral motion along a oating axis perpendicular to both the femoral exion axis and long axis of the tibia 21 . 3.2 Patellofemoral Kinematics. Patellofemoral exionextension proved to be sensitive to friction Figs. 6 and 7 . Peak patellofemoral exion occurred at the maximum exion achieved by the model and experiment. For the range of the coefcient of friction studied, the peak patellofemoral exion varied by nearly 7 deg A range of approximately 27 deg was measured in the experiment, while the nite element model predictions ranged from 26 deg to 34 deg. The model and experimental tilt and spin were both less than 10 deg Fig. 6 . Although the trend was captured for the internal tilt prole, the model underpredicted the experimental data during the swing phase, resulting in a rms error of 2.2 deg.

Fig. 5 Tibiofemoral anterior --posterior A-P experimental and model-predicted displacement as a function of the gait cycle. Plotted data include rigid body analysis with various friction values, and a deformable analysis with a coefcient of friction= 0.08 tibio- and patellofemoral. Very close agreement between the rigid body and deformable analyses allow the results to be representative of either analysis for the presented coefcient values.

Fig. 7 Experimental and model-predicted patellofemoral exion --extension as a function of the gait cycle using selected coefcient of friction values

081010-4 / Vol. 132, AUGUST 2010

Transactions of the ASME

Downloaded 27 Jul 2011 to 147.46.230.42. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

Fig. 8 Experimental and model-predicted patellofemoral translations as a function of the gait cycle + translations are superior, anterior, and lateral

The predicted patellar spin captured the trend of the experimental data, with a rms error of 1.4 deg. The coefcient of friction had little effect on the tilt and spin rotational kinematic data. The range of inferior-superior translation of the patella with respect to the femur was approximately 34 mm for both experimental and model data, with a rms error of 1.86 mm Fig. 8 . Ranges of approximately 13.5 mm for both experimental and model data were realized for the posterior translation of the patella, resulting in a rms error of less than 1.0 mm. The range of medial-lateral shift was 6.5 mm for the experimental data, while the model predicted a range of 5.0 mm, and the resulting rms error was 0.74 mm. All peak translations occurred at the maximum exion achieved by the model and the experiment. The coefcient of friction between the patella and the femur had only a small effect less than 1 mm on these translational results.

Discussion

During the design phase of implant development, parametric evaluation of the geometric variables, implant alignment, or loading conditions, and their effect on joint kinematics and contact stresses, are of great interest to optimize the potential performance. Although physical simulators are required to verify or validate the prototype designs, comprehensive evaluation is not feasible due to the time required. Veried computational tools can make an invaluable contribution during the design phase due to the efciency in studying the effects of variable perturbations. In this study, experimental kinematic analysis was performed with a semiconstrained, xed-bearing implant, and six degrees-offreedom motion was described and compared with nite element model predictions. The xtured analysis under controlled loading conditions was intentionally chosen to assess the ability of the computational methods to predict the interface motion, and to avoid the uncertainty present in the modeling of cadaveric specimens, specically the unknown ligament pretensions and properties. Substantial uncertainty is present in modeling these parameters and results in large variability in the kinematic predictions, obscuring the assessment of the model to predict the interface kinematics. Model verication with cadaveric studies should only be undertaken once this study is completed, and subsequent probabilistic studies may be completed including uncertainty in the ligamentous properties and pretension. In general, good agreement was found between the experimental and model-predicted kinematics for both rigid body and deformable representations. Translational degrees of freedom averaged approximately 1 mm of the rms error. Rms error for rotational degrees of freedom averaged near 2 deg. All experiment translational and rotational trends were captured with the model. Journal of Biomechanical Engineering

Maximum rotations and translations occurred in the experiment and the model at the deepest exion the knee achieved with the only exception being tibiofemoral internal-external rotation, which occurred near the highest internal-external torque and adduction-abduction force applications. Patellar exion and tibiofemoral anterior-posterior motions were shown to be sensitive to the interface friction, but the other degrees of freedom were generally not sensitive to friction. The tibiofemoral anteriorposterior motion component increased with a higher friction coefcient. This was likely in part due to the increase in friction, creating an increased relative rolling component while decreasing the relative sliding, thus creating greater posterior femoral relative translation greater femoral rollback . There are a number of sources of potential error in comparing the experimental and predicted motions. Due to the complexity of measuring and including friction and damping parameters, such as at the machine joints and bearings, the model included only losses at the joints necessary to minimize an oscillation seen in the predicted I-E motion near peak exion. While some of these parameters will undoubtedly affect the numerical predictions, the trends and relative magnitudes of the experiment were adequately captured with a simplied setup, including only damping at the ankle sled. While machine part masses were determined from the manufactured parts, center of mass calculations were performed on CAD models of the geometry. In addition, the Kevlar band of the extensor mechanism was, for computational efciency, represented using several spring elements. The frictional characteristics at the interface were approximated with simple Coulomb friction. For a typical setup with the Optotrak system used in the motion capture, a 10 deg rotation showed a bias of 0.05 deg and a 95% repeatability limit of 0.67 deg. A 10-mm translation showed a bias of 0.03 mm and a 95% repeatability limit of 0.29 mm 25 . A potentially more signicant source of error is the digitization of the anatomical landmarks used to develop the anatomical coordinate frames on the tibia, femur, and patella. Although not as uncertain as determining landmark data on cadaveric specimens, error in establishing the anatomical frames creates signicant variability in reported kinematics 26 . The ability to select either deformable or rigid body analysis is important to maximize analysis efciency. For the current loading conditions, the rigid body analysis was a good approximation of the kinematics predicted with a fully deformable analysis. The current study is in support of the previous verication, demonstrating that the nonlinear pressure-overclosure rigid body estimations of joint mechanics reect the results of a fully deformable model 22 . Given the conservative estimate of a stable time increment, the deformable model run time was nearly a day on a AUGUST 2010, Vol. 132 / 081010-5

Downloaded 27 Jul 2011 to 147.46.230.42. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

3.0 GHz Pentium 4 , while the corresponding rigid body analysis time was reduced to approximately 15 min. The ability to quickly predict kinematics and contact mechanics is an invaluable tool for setting up complex analyses, and reduced analysis time also allows for potential probabilistic evaluation. A probabilistic study would lend insight into the setup variables that most signicantly affect the experimental results, a task that would be very difcult to perform experimentally. The xtured analysis presented represents a controlled environment to troubleshoot and understand the limitations of the corresponding nite element model. To be effectively used in the design phase, the ability of the model to differentiate and properly rank implant designs and loading conditions needs to be veried in an expanded experimental study with multiple implant geometries and loading conditions. In addition, further verication of the kinematic prediction with implanted cadaveric knees is ongoing. The nite element model presented, including simultaneous tibio- and patellofemoral articulations, represents an important step in efcient exploration of clinically relevant parameters and their effects on the joint performance. To the authors knowledge, the results presented describe the rst experimental kinematic validation of a nite element total knee replacement model during concurrent, dynamic tibio- and patellofemoral loading.

Acknowledgment
This research was supported in part by DePuy, a Johnson & Johnson Co.

References
1 DesJardins, J. D., Walker, P. S., Haider, H., and Perry, J., 2000, The Use of a Force-Controlled Dynamic Knee Simulator to Quantify the Mechanical Performance of Total Knee Replacement Designs During Functional Activity, J. Biomech., 33, pp. 12311242. 2 Walker, P. S., Blunn, G. W., Broome, D. R., Perry, J., Watkin, S. A., Sathasivam, S., Dewar, M. E., and Paul, J. P., 1997, A Knee Simulating Machine for Performance Evaluation of Total Knee Replacements, J. Biomech., 30, pp. 8389. 3 Zavatsky, A. B., 1997, A Kinematic-Freedom Analysis of a Flexed-KneeStance Testing Rig, J. Biomech., 30, pp. 277280. 4 Zachman, N. J., Hillberry, B. M., and Kettlekamp, D. B., 1978, Design of a Load Simulator for the Dynamic Evaluation of Prosthetic Knee Joints, ASME Paper No. 78-DET-59, pp. 111. 5 Chao, E. Y. S., MacWilliams, B. A., Chan, B., and Meija, L., 1994, Evaluation of a Dynamic Joint Simulator, ASME Bioeng. Div. Publ. Bed., Vol. 28, pp. 281282. 6 Maletsky, L. P., and Hillberry, B. M., 2005, Simulating Dynamic Activities Using a Five-Axis Knee Simulator, ASME J. Biomech. Eng., 127, pp. 123 133. 7 Sathasivam, S., and Walker, P. S., 1997, A Computer Model With Surface Friction for the Prediction of Total Knee Kinematics, J. Biomech., 30, pp. 177184.

8 Sathasivam, S., and Walker, P. S., 1998, A Computer Model to Predict Subsurface Damage in Tibial Inserts of Total Knees, J. Orthop. Res., 16, pp. 564571. 9 Rawlinson, J. J., and Bartel, D. L., 2002, Flat Medial-Lateral Conformity in Total Knee Replacements Does Not Minimize Contact Stresses, J. Biomech., 35, pp. 2734. 10 Giddings, V. L., Kurtz, S. M., and Edidin, A. A., 2001, Total Knee Replacement Polyethylene Stresses During Loading in a Knee Simulator, ASME J. Tribol., 123, pp. 842847. 11 Piazza, S. J., and Delp, S. L., 2001, Three-Dimensional Dynamic Simulation of Total Knee Replacement Motion During a Step-Up Task, ASME J. Biomech. Eng., 123, pp. 599606. 12 Godest, A. C., Beaugonin, M., Haug, E., Taylor, M., and Gregson, P. J., 2002, Simulation of a Knee Joint Replacement During Gait Cycle Using Explicit Finite Element Analysis, J. Biomech., 35, pp. 267275. 13 Halloran, J. P., Petrella, A. J., and Rullkoetter, P. J., 2005a, Explicit Finite Element Modeling of Total Knee Replacement Mechanics, J. Biomech., 38, pp. 323331. 14 Miller, M. C., Zhang, A. X., Petrella, A. J., Berger, R. A., and Rubash, H. E., 2001, The Effect of Component Placement on Knee Kinetics After Arthroplasty With an Unconstrained Prosthesis, J. Orthop. Res., 19, pp. 614620. 15 Rhoads, D. D., Noble, P. C., Reuben, J. D., Mahoney, O., and Tullow, H. S., 1990, The Effect of Femoral Component Position on Patellar Tracking After Total Knee Arthroplasty, Clin. Orthop. Relat. Res., 260, pp. 4351. 16 Singerman, R., White, C., and Davy, D. T., 1995, Reduction of Patellofemoral Contact Forces Following Anterior Displacement of the Tibial Tubercle, J. Orthop. Res., 13, pp. 279285. 17 Singerman, R., Pagan, H. D., Peyser, A. B., and Goldberg, V. M., 1997, Effect of Femoral Component Rotation and Patellar Design on Patellar Forces, Clin. Orthop. Relat. Res., 334, pp. 345353. 18 Yoshii, I., Whiteside, L. A., and Anouchi, Y. S., 1992, The Effect of Patellar Button Placement and Femoral Component Design on Patellar Tracking in Total Knee Arthroplasty, Clin. Orthop. Relat. Res., 275, pp. 211219. 19 Anouchi, Y. S., Whiteside, L. A., Kaiser, A. D., and Milliano, M. T., 1993, The Effects of Axial Rotational Alignment of the Femoral Component on Knee Stability and Patellar Tracking in Total Knee Arthroplasty Demonstrated on Autopsy Specimens, Clin. Orthop. Relat. Res., 287, pp. 170177. 20 Guess, T. M., and Maletsky, L. P., 2005, Computational Modeling of a Dynamic Knee Simulator for Reproduction of Knee Loading, ASME J. Biomech. Eng., 127, pp. 12161221. 21 Grood, E. S., and Suntay, W. J., 1983, A Joint Coordinate System for the Clinical Description of Three-Dimensional Motions: Application to the Knee, ASME J. Biomech. Eng., 105, pp. 136144. 22 Halloran, J. P., Easley, S. K., Petrella, A. J., and Rullkoetter, P. J., 2005b, Comparison of Deformable and Elastic Foundation Finite Element Simulations for Predicting Knee Replacement Mechanics, ASME J. Biomech. Eng., 127, pp. 813818. 23 DeHeer, D. C., and Hillberry, B. M., 1992, The Effect of Thickness and Nonlinear Material Behavior on Contact Stresses in Polyethylene Tibial Components, Trans. Annu. Meet. - Orthop. Res. Soc., 17, p. 793. 24 Lafortune, M. A., 1984, The Use of Intra-Cortical Pins to Measure the Motion of the Knee Joint During Walking, Ph.D. thesis, Penn State University, University Park, PA. 25 Maletsky, L. P., Sun, J., and Morton, N. A., 2007, Accuracy of an Optical Active-Marker System to Track the Relative Motion of Rigid Bodies, J. Biomech., 40, pp. 682685. 26 Morton, M. A., Maletsky, L. P., Pal, S., and Laz, P. J., 2007, Effect of Variability in Anatomical Landmark Location on Knee Kinematic Description, J. Orthop. Res., 25, pp. 12211230.

081010-6 / Vol. 132, AUGUST 2010

Transactions of the ASME

Downloaded 27 Jul 2011 to 147.46.230.42. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

You might also like