Purpose of the study: Several authors have focused attention on the kinematics of total knee arthroplasty. In vitro studies have shown the influence of prosthetic kinematics on wear of the tibial insert. In vivo, investigations are limited for this important parameter. We propose computer algorithm which gives a linear assessment of femur penetration into the tibial insert. The purpose of this work was to test this measurement by radiostereometric analysis (RSA) on pre-worn prosthesis models.

Material and methods: A resin plateau was manufactured with four Triatlhon posterior stabilized inserts. Three of these inserts were pre-worn on the medial glenoid; wear varied from 0.8 to 1.2 mm. The wear pattern was controlled by a computer program in order to localize it on a posteromedial gliding band posteriorly to the posterior stabilization implant. A femoral component was implanted on dry bone and loaded on a simulator. RSA images were obtained for each of the plateau. Five series of images were obtained from 0 to 40° flexion. The measurement method consisted in defined a perfect middle plane based on the analysis of tantalum beads embedded in the resin plateau, the to define the point the closest to this plane belonging to the medial condyle. The norm of the normal vector for this plane passing through the lowest point of the femur was the reference. The difference of the norms of the vectors constructed from the healthy plateau and the worn plateaus was defined as penetration of the femur into the tibia insert. The rood mean square (RMS) method was used to measure intraobserver variability and the interclass coefficient of correlation (ICC) was determined.

Results and Discussion: Our results were encouraging. 0.8/to 1.12 mm wear was detected with very good accuracy. The data spread was however wide, irrespective of the wear value. The images in flexion of this model provided the most reliable readings. This study is the first offering an approach to wear measurement in vivo. The data spread should be improved by fine tuning the experimental model and the pre-wear pattern.

Conclusion: In vivo measurement of TKA wear remains a challenge in clinical medicine. Our work demonstrate the feasibility of using RSA.

Purpose of the study: Comprehension of total knee arthroplasty (TKA) kinematics is primordial for improving the functional outcome and longevity of these prostheses. Several methods are available for evaluating knee kinematics. The purpose of this study was to determine the accuracy of the 2D fluoroscopic method in vitro, taking optoelectronic analysis as the gold standard.

Material and methods: In order to compare these two techniques, a posterior stabilised prosthesis was implanted on dry bones. The lateral ligaments were modellised with two elastic bands. Thirty flexion movements were imposed consecutively. The kinematics of this prosthetic model were recorded simultaneously using the fluoroscope and a computer-assisted surgery system. The technique used for the fluoroscopic analysis was based on the detection of the contours and projective geometry algorithms. The statistical analysis measured differences and correlations between the two systems using the root mean square (RMS) method and interclass coefficients of correlation (ICC) in addition to Bland and Altman analyses.

Results: Three hundred thirty six relative implant positions were analysed for 30 flexions from −8 to 132 degrees. The objective RMS were to the order of one degree for flexion, varus and tibia rotation. Conversely, there was a difference of 2.43±3.17 mm for the mediolateral distance (ML). Similarly the ICC were to the order of 0.9 for the six degrees of freedom of the model with the exception of ML displacement where the ICC was 0.106. These analyses were confirmed by the Bland and Altman analysis which revealed an underestimation of the ML distance by the fluoroscopic method in greatest internal rotation.

Discussion: This study is the first using a realistic model to evaluate the kinematic data provided by 2D fluoroscopy in comparison with conventional navigation data. The results show a good agreement between the two techniques and a small difference in measures excepting for the ML plane. The results are less satisfactory than those reported in the literature where data were obtained from computer simulations.

Conclusion: 2D fluoroscopy of the TKA kinematics provides precise data. Nevertheless, the limits and inaccuracies of this technique should be recognized. This study is a prerequisite for in vivo 2D fluoroscopy.