I thank Dr. John for his interest in our work involving in-vivo total knee arthroplasty (TKA) contact kinematics(1). His questions and comments are important and timely. We believe that a scientific discussion will definitely help readers of the Journal better understand in-vivo knee joint kinematics, especially tibiofemoral articular contact kinematics.

It is true that the term “posterior femoral translation” is not synonymous with “femoral rollback.” We use “posterior femoral translation” exclusively when referring to our own data on femoral condylar motion or contact point motion. However, we also noticed in the literature that a precise description of the tibiofemoral kinematics is not a trivial task, especially when the medial and lateral compartments are involved. “Femoral rollback” has been sometimes used as a "short cut" term when describing “posterior femoral translation”. If any of these terms are used, a clear definition has to be given.

Dr. John is correct in his reservations about using a nearest point methodology to determine point of contact on articular surfaces, especially when attempting to report contact between conforming articulating surfaces. This method has been used in previous studies and has provided much of the initial data on knee kinematics. We have published an article in the Journal of Biomechanics that compared the contact locations determined using the nearest point methodology with contact determined using intersecting surfaces in human knee joints [2]. This comparison indicated that articular contact kinematics are better measured from the intersection of the articulating surfaces. When determining the tibiofemoral contact of a cruciate-retaining TKA in our study [1], the intersection of the articular surfaces was measured for both the medial and lateral compartments. The resulting intersection between the femoral component and the proximal tibial polyethylene insert produced an area of contact. As described in the Materials and Methods section, the area centroid was calculated and used to compare the locations of the contact areas at each flexion angle. Therefore, the contact points reported in our study represent the center of the contact areas. We agree that this contact point may not be the location of peak pressure. However, this data analysis can provide a quantitative and consistent way to report contact kinematics. Peak contact location can only be obtained through a 3D finite element calculation in our cases.

In our current manuscript [1], we did a general comparison of the contact kinematics of patients after cruciate retaining TKA to the cartilage contact kinematics of normal, healthy subjects [2,3]. Dr John is right that if the data are to be compared quantitatively, the same reference should be used. In our work, we plotted the contact points directly on the tibial plateau surfaces to demonstrate the actual contact locations. Therefore, qualitative comparison can be done, as discussed in our paper.

References:

1. Li et al. Three-dimensional articular contact knee kinematics of a cruciate-retaining TKA. JBJS 2006; 88-A(2): 395-402.

2. DeFrate et al. In vivo tibiofemoral contact analysis using 3D MRI- based knee models. J Biomech. 2004; 37(1):1499-504.

3.Li et al. In vivo articular cartilage contact kinematics of the knee: an investigation using dual-orthogonal fluoroscopy and magnetic resonance image-based computer models. Am J Sports Med. 2005; 33(1): 102- 7.

To The Editor:

I read with interest the article by Li, et al, (1). The paper confirms that it is not possible to have asymmetric posterior translation of the lateral and medial femoral condyles without some mediolateral translation. However, the paper raises a few questions.

There seems to be a tendency to use posterior translation of tibiofemoral contact synonymously with femoral roll-back. Femoral roll back is the posterior translation of the variable instant centres of rotation of the respective femoral condyles, especially the lateral condyle. This would certainly correlate and probably mirror the posterior translation of tibiofemoral contact points but only in a stable knee. In an unstable knee with the loss of the ACL, where anteroposterior tibiofemoral translation is possible even without flexion, to attribute any posterior translation of tibiofemoral contact point to the changes in instant centres of rotation of the femoral condyles and hence flexion of knee alone would certainly seem misplaced. Even if it were to be argued that posterior translation would be governed by the presence of the PCL alone, the initial contact point would certainly be anterior in the absence of the ACL.

There seems to be a huge disparity in the methodology used for measuring femoral rollback. Patil, et al, (2) reported results in unicompartmental knees using the midpoint of the transepicondylar line as the reference point to measure femoral roll-back. The authors have used the nearest points between the metal tibial tray and femoral component to deduce the point of contact of the components with the plastic tray. This methodology would be feasible if the polyethylene were flat as some of the PCL retaining designs were. However,in the presence of a dished polyethylene component as illustrated in the figures presented in this article,the accuracy of their technique would be questionable since there is no point of contact anymore, but only an area of contact. It would seem that in the presence of conforming articulating surfaces, the geometry of the surfaces would bear a big influence on the closest points between the femoral and tibial components and that a specific point may not necessarily represent the point of maximum compression. The observation that the posterior lip of the polyethylene tray impinges at the end of flexion may possibly be due to this.

The authors have mentioned using the methodology in normal patients and have reported a similar kinematic profile in them, but they have not elaborated on how the contact points were determined. They have not defined the stationary point from which posterior translation was defined. In contrast, in knees with a prosthesis implanted, specific points on the metal tray can be identified to allow for reproducible measurements.

I believe that validation of methodolgy with observer errors is important. It is also important not to ascribe all posterior tibiofemoral translation to have been caused secondary to changes in instant centres of rotation of the femoral condyles (Femoral roll-back) and hence flexion at the knee, although I agree that differentiation of the individual contributions may be difficult if not entirely impossible.

It would be very helpful to readers of the Journal if there were a consensus on terms and methodology used in measuring tibiofemoral kinematics.

References:

1. Li, et al. Three-Dimensional Tibiofemoral Articular Contact Kinematics of a Cruciate-Retaining Total Knee Arthroplasty. J Bone Joint Surg Am. 2006; 88:395-402

2. Patil, et al. Can Normal Knee Kinematics Be Restored with Unicompartmental Knee Replacement? J Bone Joint Surg Am. 2005;87: 332-338.