Aims

The objective of this study is to assess the use of ultrasound (US) as a radiation-free imaging modality to reconstruct 3D anatomy of the knee for use in preoperative templating in knee arthroplasty.

At present, long-term follow-up studies are used to assess the performance and longevity of an implant, but the downside is that designers must wait 5–10 years before they receive this feedback. Therefore, the objective of this study was to develop a theoretical simulator that will allow for prediction of kinematic patterns based on implant shape and prediction of implant longevity based on the implant’s ability to adapt to in vivo conditions.

A model of the normal lower leg, including muscles and all ligament structures, was developed using Kane’s theory of dynamics. All muscles and ligaments were modeled as distributed loads and included wrapping points to follow the true path of soft-tissue structures.

Currently, two activities are available to the user: leg extension and deep flexion. 3D shapes, pertaining to the implant designs are input to the model.

A validation of the model was conducted using an initial force prediction for each muscle. The predicted kinematics were compared to a library of in vivo kinematics from over 2000 knees obtained using fluoroscopy and a 3-D model fitting technique. If the kinematic patterns from the model were incorrect, an optimization feedback algorithm induced a change in the muscle force. This process continued until the proper muscle force profiles were determined.

Then, using muscle forces which achieve observed motion in TKA previously implanted and analyzed, evaluation of various new implant designs could be assessed.

Altering designs or constraints in TKA lead to quite different kinematic profiles, even when the same muscle force profiles are used. Further research needs to be conducted using more design profiles before multiple implant designs could be evaluated and compared.

Conventional fluoroscopes are routinely used to analyze human skeletal joints during motions such as deep knee bends. Such diagnostics are used to characterize pre and post operative arthoplasty results, particularly in association with total joint replacement procedures. The pseudo-stationary conditions imposed by the fixed fluoroscope limit the diagnostic procedures to much less than natural skeletal motion and load conditions, thus diminishing the utility of the results. A new class of fluoroscopy has been developed in which a robotic mechanization is used to allow selected joints to be x-rayed while the human subjects perform natural motions such as walking. The tracking fluoroscope system (TFS) is essentially a mobile robot that can acquire real-time x-ray records of hip, knee, or ankle joint motion while the patient walks normally within a laboratory floor area. It is anticipated that the TFS will provide clearer and more representative x-ray images.

The robotic mechanization includes an untethered and omni-directional mobile platform that follows the patient as he/she walks, including negotiating stairs or ramps.

In addition to following the patient, additional control devices track the joint motions that occur relative to the patient’s body, e.g., knee joint vertical and anterior/posterior relative motion. The technical features of the TFS will be described, and test results related to the commissioning of the TFS for clinical trials will be presented. Initial clinical test results will be provided.

An institution of the authors (Center for Musculoskeletal Research) and one author (DAD) have received funding from DePuy, Inc. (Warsaw, IN).

Each author certifies that his or her institution has approved the reporting of these cases, that all investigations were conducted in conformity with ethical principles of research, and that informed consent for participation in the study was obtained.

This work was performed at Center for Musculoskeletal Research, University of Tennessee, Knoxville, TN and the Rocky Mountain Musculoskeletal Research Laboratory, Denver, CO.

Previous clinical studies have documented the incidence of squeaking in subjects having a ceramic-onceramic (COC) THA. An in vivo sound sensor was recently developed used to capture sound at the THA interface. In this first study, it was determined that subjects having all bearing surface types demonstrated variable sounds. Therefore, in this follow-up study, the overall objective was to simultaneously capture in vivo sound and motion of the femoral head within the acetabular cup during weight-bearing activities for subjects implanted with one of four different ceramic-on-ceramic (COC) THA.

Twenty subjects, each implanted with one of four types of Ceramic-on-Ceramic THA (9 Smith and Nephew, 8 Stryker, 2 Wright Medical Technologies and 1 Encore) were analyzed under in vivo, weightbearing conditions using video fluoroscopy and a sound sensor while performing gait on a treadmill. Patients were pre-screened and two groups were defined: a group diagnosed as audible squeakers (9 THAs) and a control group of THA patients not experiencing audible sounds (11 THAs). Two tri-axial piezoelectric accelerometers were attached to the pelvis and the femoral bone prominences respectively. The sensors detect frequencies propagating through the hip joint interaction. Also, 3D kinematics of the hip joint was determined, with the help of a previously published 2D-to-3D registration technique. In vivo sound was then correlated to 3D in vivo kinematics to determine if positioning of the femoral head within the acetabular cup is an influencing factor.

For the audible group, two had a Smith and Nephew (S& N) THA, six a Stryker THA and one a Wright Medical (WMT) THA. Both of the S& N subjects, 5/6 Stryker and the Wright Medical subjects experienced femoral head separation. The maximum separation for those subjects was 4.6, 5.0 and 2.1 mm for the S& N, Stryker and WMT subjects, respectively. The average separation was 4.3, 2.0 and 2.1 mm for the S& N, Stryker and WMT subjects, respectively. For the eleven subjects in the control group, seven subjects had a S& N THA, two a Stryker and one each having a WMT and Encore THA. All 11 of these subjects demonstrated hip separation with the maximum values being 3.8, 3.4, 1.9 and 2.4 mm for the S& N, Stryker, WMT and Encore THA, respectively. The average separation values were 1.8, 2.3, 1.9 and 2.4 mm for the S& N, Stryker, WMT and Encore THA subjects, respectively.

Four distinct sounds were produced by subjects in this study, which were squeaking, knocking, clicking and grating. Only 3/20 subjects produced a “squeaking” sound that was detected using our sound sensor. One of these subjects had a Stryker THA and two had a WMT THA. Further analysis of the nine subjects who were categorized as audible squeakers revealed that only 0/2, 1/6 and 1/1 subjects having a S& N, Stryker and WMT THA, respectively, demonstrated a squeaking sound that was detected using our sound sensor. Both (2/2) S& N subjects demonstrated a knocking and clicking sound, but neither produced a grating sound, while 5/6 Stryker subjects produced a knocking sound, but only 1/6 demonstrated a clicking or grating sound. Besides the squeaking sound, the only other sound produced by the WMT audible squeaker was a knocking sound. Only 1/11 control group subjects demonstrated a squeaking sound, which was a subject having a WMT THA. With respect to the control group subjects having a S& N THA, 5/7, 1/7 and 3/7 subjects produced a knocking, clicking or grating sound, respectively. Only 1/2 subjects having a Stryker THA produced a knocking or grating sound.

This is the first study to compare multiple COC THAs in analyzing correlation of femoral head separation (sliding) and sound. It was seen that all the THA groups had occurrences of separation and each case of separation correlated with the sound data. These results lead the authors to believe that the influence of squeaking is multi-factorial, and not necessarily attributed only to the bearing surface material.

Previously, Komistek et al have demonstrated anomalous behaviours in total joints such as separation (sliding) in THAs and condylar lift-off in TKAs. These cases result in reduced contact area, increased contact pressure, polyethylene wear and could induce prosthetic loosening and joint instability.

However, here is no known research done on correlating kinematic conditions with acoustic data for the tibio-femoral joint interface. This study deals with the development of a new method to diagnose such conditions using sound and frequency data. The objective of this study was to determine and compare the in vivo, 3D kinematics and sound for younger subjects with a normal knee, to those of older subjects, with an unimplanted and implanted knee joint.

Ten older subjects having a Hi-Flex PS MB TKA and a contralateral non implanted knee and five younger subjects (with a normal knee) were analysed under in vivo, weight-bearing conditions using video fluoroscopy and a sound sensor while performing four different activities.

deep knee bend to maximum flexion

Three piezoelectric triaxial accelerometers were attached to the femoral epicondyle, tibial tuberocity and the patella respectively. The sensor detects frequencies that are propagated through the tibio-femoral interaction. The signal from the accelerometers was then transferred to a signal conditioner for signal amplification. A data acquisition system was then connected to receive the amplified signal from the signal conditioner and transfer it to a laptop for storage. A sampling rate of 10500Hz was used and frequencies upto 5000Hz were recorded. The signal was then converted to audible sound. Also, 3D tibio-femoral kinematics of the knee was determined, for the four activities with the help of a previously published 2D-to-3D registration technique. The fluoroscopy video and the sensor measurements were synchronized, analysed and compared from full extension to maximum knee flexion for DKB, one full cycle of gait, one complete step on stair climb and from sit-to-stand positions in chair rise.

On average the subjects achieved more flexion with their TKA than with their contralateral knee and consequently experienced significantly higher ROM for their implanted knee. However, both of these groups achieved lower ROM than the normal knees. Significant differences were seen in the AP position of the tibiofemoral contact point. The contact point of the medial condyle for the TKA knee was significantly more posterior at 0° and 30° and remained more posterior than the same condyle of the contralateral throughout flexion. Posterior femoral rollback was seen in all groups, with the normal knee achieving significantly higher posterior femoral rollback when compared to the contralateral and TKA knees. Audible signals were observed for all three groups of knees. The frequency analysis revealed that specific frequencies for all groups were within the same range, but the most dominant frequency for each varied. This may be related to the variable interaction surfaces leading to different dominant frequencies which were excited at magnitudes related to the type and condition of material being impacted (polyethylene/meniscus).

This was the first study to correlate in vivo kinematics to in vivo sounds in the knee. The sounds that were detected correlated well to in vivo motions, especially abnormal kinematic patterns. The ultimate aim of this study is to create a stand alone tool (based only on sound data) that could be used as a diagnostic tool to determine total joint conditions and reduce the dependence on radiation techniques.

Polyethylene debris can cause patient osteolysis, patient pain and discomfort, and implant revision. Previous fluoroscopic studies have determined the incidence of femoral head separation from the acetabular cup, but clinical significance of this phenomenon has not been established. It has been hypothesized that hip separation may lead to polyethylene wear, while others hypothesize that hip separation may be occurring due to wear. Therefore, the purpose of the study is to conduct an in vivo kinematic analysis to determine if there is a correlation between-femoral head separation and wear and to utilize a mathematical modeling to determine the clinical significance of these variables.

Twenty subjects were strategically selected to participate in this study. Ten subjects were determined to have at least1.0 mm of polyethylene wear, while ten subjects had less than 0.1 mm of polyethylene wear. All 20 patients were asked to perform gait on a treadmill while under fluoroscopic surveillance. The incidence of femoral head separation was determined for each subject. Then, a three-dimensional mathematical model of the hip joint was used to determine bearing surface conditions for each subject.

Fifty-five percent of the subjects evaluated demonstrated femoral head separation. Subjects deemed to have greater than 1.0 mm of wear experienced less separation, on average and overall magnitude than subjects without wear. In this study, only 10% of the subjects tested-demonstrated no wear and no separation. The derived force profiles in this study were greater for both groups, compared with the non-implanted hips, previously evaluated. The forces in the hip joint ranged from 2.0 to 3.0times body weight.

Although it was expected that subjects having more wear would have greater magnitudes of femoral head separation, the opposite was true. Further kinetic analysis determined that the subjects having wear also experienced greater force profiles through gait. Therefore, it is assumed that the subjects having wear may have been-implanted with a tighter socket, thus leading to greater shear forces.

Previous studies of cervical fusion have noted the appearance of new degeneration at levels adjacent to fused segments. The cause of this degeneration has not been accurately determined. The objective of this project is to determine the forces in normal and fused spines in vivo and compare the results to see if the forces in the fused spine are increased enough to cause degeneration in adjacent levels.

A normal and a fused patient (one level fused in C5-C6) have been chosen to perform a full flexion and extension motion experiments. Kinematic functions were obtained from the images. Data was input into the mathematical model and the kinetic results have been decided.

The result can help us understand in vivo kinematical and kinetic characteristics of cervical spine fusion and develop further studies in 3D models. The calculated forces will be compared to previously reported data to determine if the expected increased forces in the fused spinal are great enough to cause degeneration of adjacent levels. A better understanding will help in better treatment of cervical spine disorders.

Previously, we conducted a single surgeon in vivo kinematic study that revealed subjects having a PCR TKA with asymmetrical condyles experienced a high incidence of posterior femoral rollback. Therefore, the objective of this follow-up study was to determine if posterior femoral rollback from our single surgeon series can be attributed to the implant design, surgical technique, or the presence of a well functioning PCL.

Three-dimensional femorotibial contact positions for eighty subjects, implanted by three surgeons, were evaluated using fluoroscopy during a deep knee bend. Twenty subjects had a PCR TKA without a PCL, while the other 60 subjects were deemed to have a functional PCL. All subjects were implanted with a PCR TKA having a larger lateral radius of curvature compared to the medial condyle.

Fifty-four of sixty subjects in this study having a functional PCL experienced posterior femoral rollback of the lateral condyle, while 13/20 subjects not having a PCL experienced posterior femoral rollback. Also, 48/60 subjects having a well functional PCL and 10/20 subjects without a PCL experienced a normal axial rotation pattern. The incidence for condylar lift-off was low, and only 6/80 subjects in this study experienced greater than 2.0 mm of condylar lift-off. Forty of sixty subjects having a well-functional PCL experienced greater than 100 degrees of weight-bearing range-of-motion and the maximum weight-bearing range-of-motion was 144o.

The subjects in this study experienced excellent kinematic patterns, consistent to the normal knee, although less in magnitude than the normal knee. Surprisingly, on average, subjects in this study without a PCL experienced posterior femoral rollback of the lateral condyle leading to the assumption that the PCL did not play a significant role in the excellent kinematic patterns achieved by subjects in this study. The results from this study do support the hypothesis derived from our single surgeon series that asymmetric condylar geometry may lead to better kinematic patters for subjects implanted with a fixed bearing PCRTKA, as it appears implant design is the most influencing factor that lead to normal kinematic patterns.

The objective of this study was to compute the in vivo dynamic tibiofemoral contact forces for normal alignment, and then evaluate the change in contact forces and pressures with increasing varus-valgus and internal-external rotational malalignment of the femoral component. A three-dimensional computational model of the lower limb during deep knee bend was created using Kane’s method of dynamics. The change in forces from normal with malalignment of up to 10° valgus, 10° varus, 10° internal axial femoral rotation, and 10° internal axial femoral rotation were determined. In this study, varus-valgus malalignment had the greatest effect on medial-lateral pattelofemoral contact forces, with a maximum increase of 2.25 times body weight for 10° valgus malalignment. Axial malalignment had the greatest influence on tibiofemoral contact forces.

At present, computational modeling has not been utilized as a design tool for total knee replacement (TKR). Also, classifying a new design as successful usually requires many years of long-term clinical follow-up studies. Computational modeling presents an opportunity to contribute to implant design evaluations and prediction of long-term success, during the early stages of the implant design process. The purpose of this study was to construct a computational model that will determine and compare in vivo dynamic forces and torques of the non implanted and implanted knees. It is hypothesized that this model will provide valuable information pertaining to post-implantation boundary conditions during the design phase.

A three-dimensional (3-D), inverse dynamics model of the human lower limb was created. System differential equations were derived for the human lower extremity using Kane’s theory of dynamics.Input kinematics were obtained for five normal knees and five posterior stabilized TKR, determined while subjects performed deep knee bend while under fluoroscopic surveillance. Musculo tendinous units were assumed to act along straight line segments, and ligamentous units were represented by nonlinear elastic elements. Knee kinetics were calculated and compared fo reach group and a comparison was conducted.

Kinetics were much more variable for the TKR group, and tibiofemoral contact forces were on average higher than the normal group: 2.47 times body weight (BW) and 2.21 BW, respectively. Increased posterior femoral rollback lead to lower axial contact forces and lower quadriceps forces in both groups. Force patterns were very sensitive to input patient specific kinematics.

The predicted tibio femoral forces were higher in TKR subjects, which is consistent with current clinical knowledge. Force patterns for the normal subjects were more consistent than those forthe TKR subjects, which was primarily attributed to the greater variance in kinematics for the TKR subjects. This study represents a first step in constructing a design facilitation tool for TKR technology. Successful designs will be determined by producing kinetic patterns most similar to normal knee patterns.

At present, contact stress analyses of TKA involve in vitro experimental testing. The objective of this project was to develop a parametric mathematical model that determines in vivo contact stresses for subjects implanted with a TKA, under in vivo, dynamic conditions. It is hypothesized that the results from this model will be more representative of in vivo conditions, thus leading to more accurate prediction of TKA bearing surface stresses.

In vivo kinematics were determined for ten subjects implanted with a posterior stabilized TKA during gait and a deep knee bend under fluoroscopic surveillance. Three-dimensional contact positions, determined between the femoral component and the polyethylene insert, were entered into a complicated mathematical model to determine bearing surface forces. In vivo kinematics and kinetics were entered into a deformation model to predict in vivo contact areas between the medial and lateral condyles and tibial insert. The orientation of the femoral and tibial components, the predicted in vivo contact areas, and vectoral information of soft-tissue derived from MRI images were then entered into a mathematical model that predicted in vivo contact stresses between the femoral component and the tibial insert.

This is the first computational model that utilizes fluoroscopy, MRI, deformation characteristics and Kane’s theory of Dynamics to predict in vivo contact stresses. Although previous models have not been validated, this model was validated by comparing the predicted foot/ ground force with the experimentally derived force. This study demonstrates that patellar motion influences forces throughout the lower extremity. The in vivo contact stress values predicted in this initial study were less than the yield strength of polyethylene.

Numerous dynamic studies have evaluated the tibiofemoral contact pressures that follow total knee arthroplasty (TKA), and several static studies utilizing finite elements and pressure sensitive film have evaluated malalignment. The objective of this study was to compute the in vivo dynamic tibiofemoral contact forces for normal alignment and evaluate the change in contact pressure with increasing malalignment of the femoral component.

A three-dimensional computational model of the lower limb during deep flexion was created using Kane’s method of dynamics. A hybrid approach was used to determine the boundary conditions of the model. The motions of a total knee arthroplasty patient were measured using fluoroscopy. The motions of the patient were varied from the normal motions to simulate malalignment of the femoral component. The change in forces with malalignments of up to 10° valgus, 10° varus, 10° internal rotation, and 10° internal rotation were determined.

An increase in the axial tibiofemoral contact force from 2.44 times body weight (BW) to 2.62 BW and a decrease in the quadriceps force from 6.8 to 5.65 BW were observed with varus malalignment. The medial-lateral patellofemoral contact force decreased from 0.95 BW to 0.1 BW with 10° varus positioning of the femur and increased to 2.2 BW with 10° valgus positioning of the femur and a decrease in the patellar ligament forces from 1.70 to 1.63 BW was observed.

Changes in the tibiofemoral and patellofemoral forces of 1–2 BW were observed as the femur was malaligned with respect to the tibia. The most significant of these changes was the medial-lateral patellofemoral contact force. The implications of these findings are that malalignment could result in increased patellar subluxation or increased wear of the polyethylene component. Concerns were raised that this initial subject evaluated may not have had optimum alignment, thus leading to more optimal bearing surface stress conditions with varus malalignment. Future studies will be evaluated for subjects having the joint line restored to conditions for non-implanted knees.

The objective was to assess and compare polyethylene-bearing mobility patterns and magnitudes in various total knee arthroplasty(TKA) types of mobile bearing TKA.

In vivo kinematics were determined for 38 subjects implanted with either a PCL-retaining (PCR) mobile bearing TKA, which allows both rotation and antero-posterior (AP) translation (n=20), aposterior stabilized rotating platform (PS) TKA (n=9) or a PCL-sacrificing (PCS) rotating platform TKA (n=9) using video fluoroscopy. Using a 3D model-fitting technique, kinematics were determined during a weight-bearing deep knee bend. The femoral and tibial components and mobile bearing polyethylene insert (implanted with four tantalum beads) were overlaid onto the fluoroscopic images to determine bearing mobility. AP bearing translation was determined for subjects implanted with a PCR mobile bearing TKA. Subjects implanted with PCR and PCS TKA were evaluated at a single interval. Those with a PS TKA were evaluated at two postoperative intervals, (12 months apart) to assess changes in bearing mobility over time.

All subjects experienced polyethylene bearing rotation relative to the tibial tray and minimal rotation relative to the femoral component. The average maximum amount of bearing rotation was 10.3o (3.0o to 20.8o), 8.9o (5.3o to 14.1o), and 8.5o (3.3o to 12.9o) for subjects implanted with a PCR, PS, and PCS mobile bearing TKA, respectively. For subjects implanted with a PS mobile bearing TKA, bearing mobility increased to 9.8o (4.8o to 14.1o) one year later post-operatively. All subject shaving a PCR mobile bearing TKA experienced AP bearing translation, averaging 5.6 mm (1.0 mm to 12.5 mm).

These results demonstrate that the polyethylene bearing is rotating and translating relative to the tibial tray in all subjects. Minimal motion occurred between the femoral component and the polyethylene insert. Magnitude and direction of bearing motion varied among subjects. Paradoxical anterior translation of the bearing during deep flexion was observed in the PCR TKA group. The presence of bearing mobility should result in lower contact stresses, reducing the potential for polyethylene wear.

Previously, fluoroscopy has been used to determine the in vivo kinematics during gait, step-up maneuvers and flexion to nine tydegrees. Recently, TKAs have been designed for deep flexion maneuvers. Therefore, the objective of this study is to determine the in vivo kinematics for subjects implanted with either a fixed or mobile bearing deep flexion TKA from full extension to maximum knee flexion

Three-dimensional femorotibial contact positions for thirty-nine subjects, implanted by two surgeons, were evaluated using fluoroscopy into deep flexion. Nineteen subjects had a fixed bearing PS deep flexion TKA and 20 subjects were implanted with a mobile bearing deep flexion TKA. Both TKA designs have similar design features, including condylar geometries.

Seventeen of nineteen subjects implanted with a fixed bearing deep flexion TKA experienced posterior femoral rollback, while all 20subjects having a mobile bearing deep flexion TKA experienced poster iorfemoral rollback. On average, subjects experienced -5.1 and -8.1 mm of posterior femoral rollback, for the fixed and mobile bearing TKA, respectively. The maximum amount of posterior femoral rollback was -11.8 and -12.4 mm for subjects having a fixed and mobile bearing TKA, respectively. On average, subjects experienced 6.5 and 5.4 degrees of normal axial rotation for a fixed and mobile bearing TKA, respectively. The average amount of weight-bearing range of motion was 116 and 125 degrees for a fixed and mobile bearing TKA, respectively. Also, subjects having both TKA types evaluated in this study experienced excellent patellofemoral kinematics

This is the first study to evaluate femorotibial and patellofemoral knee kinematics into deep flexion for a fixed and mobile bearing TKA, designed for deep flex-ion activities. Both groups in this study experienced, on average and subject-to-subject comparison, excellent kinematic patterns. Also, both TKA types evaluated in this study achieved excellent weight-bearing range-of-motion, supporting the design goal for these TKA.

The objective of this study was to determine the location of polyethylene post position and/or axis of polyethylene (PE) bearing rotation in order to maximize the rotational freedom of the PE bearing in a posterior-stabilized mobile-bearing TKA.

Kinematic data obtained in a previous study involving subjects implanted with the PFC Sigma RP (PS) was used in two mathematical models to determine the optimal configuration of the implant’s features. An inverse dynamics mathematical model used the kinematic input to calculate interactive forces between the implant components. The second mathematical model used the femur-polyethylene and polyethylene-tibial plate interactive forces in a forward solution giving the amount of polyethylene bearing rotation. Researchers altered the location of cam/post interaction and/or bearing rotation to determine the criteria for optimal bearing rotation.

During flexion, the maximum femur-polyethylene contact force calculated by the inverse model was 1.9 x BW, at maximum flexion. Maximum quadriceps, patello-femoral, and patellar ligament forces were approx. 2.9 x BW, 2.8 x BW, and 1.5 x BW at maximum flexion, respectively. We determined that the sample group experienced an average maximum bearing rotation of approximately 3.5°. Maximum bearing rotation reached approx 12.5° (10°–15°) with a 5mm lateral shift in cam/post engagement. Bearing rotation reached approximately 17.5° (15°–20°) by shifting the bearing axis 5mm posterior to that of the current design. Shifting the cam/post mechanism or bearing axis by greater than 5mm in any direction produced undesirable results.

The mathematical models used in this study were verified by comparing kinematic results obtained from a 3-D model-fitting program whereby models are matched to their respective silhouettes in a 2-D fluoroscopic image. Results from this study show that the rotational freedom of the PE bearing can be optimized by shifting its axis of rotation posterior to its present location.

We carried out weight-bearing video radiological studies on 40 patients with a total knee arthroplasty (TKA), to determine the presence and magnitude of femoral condylar lift-off. Half (20) had posterior-cruciate-retaining (PCR) and half (20) posterior-cruciate-substituting (PS) prostheses. The selected patients had successful arthroplasties with no pain or instability. Each carried out successive weight-bearing knee bends to maximum flexion, and the radiological video tapes were analysed using an interactive model-fitting technique.

Femoral lift-off was seen at some increment of knee flexion in 75% of patients (PCR TKA 70%; PS TKA 80%). The mean values for lift-off were 1.2 mm with a PCR TKA and 1.4 mm with a PS TKA. Lift-off occurred mostly laterally with the PCR TKA, and both medially and laterally with the PS TKA. Separation between the femoral condyles and the articular surface of the tibia was recorded at 0°, 30°, 60° and 90° of flexion. Femoral condylar lift-off may contribute to eccentric polyethylene wear, particularly in designs of TKA which have flatter condyles. Coronal conformity is an important consideration in the design of a TKA.