Understanding the long-term effects of total knee arthroplasty (TKA) on joint kinematics is vital to assess the success of the implant design and surgical procedure. However, while in vitro cadaveric studies quantifying post-operative biomechanics primarily reflect joint behaviour immediately after surgery,¹in vivo studies comprising of follow-up TKA patients often reflect joint behaviour a few months after surgery.² Therefore, the aim of this cadaveric study was to explore the long-term effects of TKA on tibiofemoral kinematics of a donor specimen, who had already undergone bilateral TKA, and compare them to post-operative kinematics reported in the literature.

Two fresh-frozen lower limbs from a single donor (male, age: 83yr, ht: 1.83m, wt: 86kg), who had undergone bilateral TKA (Genesis II, Smith&Nephew, Memphis, USA) 19 years prior to his demise, were obtained following ethical approval from the KU Leuven institutional board. The specimens were imaged using computed tomography (CT) and tested in a validated knee simulator³ replicating active squatting and varus-valgus laxity tests. Tibiofemoral kinematics were recorded using an optical motion capture system and compared to various studies in the literature using the same implant – experimental studies based on cadaveric specimens (CAD)^1,4 and an artificial specimen (ART)⁵, and a computational study (COM)⁶.

Maximum tibial abduction during laxity tests for the left leg (3.54°) was comparable to CAD (3.30°), while the right leg exhibited much larger joint laxity (8.52°). Both specimens exhibited valgus throughout squatting (left=2.03±0.57°, right=5.81±0.19°), with the change in tibial abduction over the range of flexion (left=1.89°, right=0.64°) comparable to literature (CAD=1.28°, COM=2.43°). The left leg was externally rotated (8.00±0.69°), while the right leg internally rotated (−15.35±1.50°), throughout squatting, with the change in tibial rotation over the range of flexion (left=2.61°, right=4.79°) comparable to literature (CAD=5.52°, COM=4.15°). Change in the femoral anteroposterior translation over the range of flexion during squatting for both specimens (left=14.88mm, right=6.76mm) was also comparable to literature (ART=13.40mm, COM=20.20mm).

Although TKA was reportedly performed at the same time on both legs of the donor by the same surgeon, there was a stark difference in their post-operative joint kinematics. A larger extent of intraoperative collateral ligament release could be one of the potential reasons for higher post-operative joint laxity in the right leg. Relative changes in post-operative tibiofemoral kinematics over the range of squatting were similar to those reported in the literature. However, differences between absolute magnitudes of joint kinematics obtained in this study and findings from the literature could be attributed to different surgeons performing TKA, with presumable variations in alignment techniques and/or patient specific instrumentation, and the slightly dissimilar ranges of knee flexion during squatting.

In conclusion, long-term kinematic effects of TKA quantified using in vitro testing were largely similar to the immediate post-operative kinematics reported in the literature; however, variation in the behaviour of two legs from the same donor suggested that intraoperative surgical alterations might have a greater effect on joint kinematics over time.

Orthopaedic training sessions, vital for surgeons to understand post-operative joint function, are primarily based on passive and subjective joint assessment. However, cadaveric knee simulators, commonly used in orthopaedic research,. 1. could potentially benefit surgical training by providing quantitative joint assessment for active functional motions. The integration of cadaveric simulators in orthopaedic training was explored with recipients of the European Knee Society Arthroplasty Travelling Fellowship visiting our institution in 2018 and 2019. The aim of the study was to introduce the fellows to the knee joint simulator to quantify the surgeon-specific impact of total knee arthroplasty (TKA) on the dynamic joint behaviour, thereby identifying potential correlations between surgical competence and post-operative biomechanical parameters. Eight fellows were assigned a fresh-frozen lower limb each to plan and perform posterior-stabilised TKA using MRI-based patient-specific instrumentation. Surgical competence was adjudged using the Objective Structured Assessment of Technical Skills (OSATS) adapted for TKA. 2. All fellows participated in the in vitro specimen testing on a validated knee simulator,. 3. which included motor tasks – passive flexion (0°-120°) and active squatting (35°-100°) – and varus-valgus laxity tests, in both the native and post-operative conditions. Tibiofemoral kinematics were recorded with an optical motion capture system and compared between native and post-operative conditions using a linear mixed model (p<0.05). The Pearson correlation test was used to assess the relationship between the OSATS scores for each surgeon and post-operative joint kinematics of the corresponding specimen (p<0.05). OSATS scores ranged from 79.6% to 100% (mean=93.1, SD=7.7). A negative correlation was observed between surgical competence and change in post-operative tibial kinematics over the entire range of motion during passive flexion – OSATS score vs. change in tibial abduction (r=−0.87; p=0.003), OSATS score vs. change in tibial rotation (r=−0.76; p=0.02). When compared to the native condition, post-operative tibial internal rotation was higher during passive flexion (p<0.05), but lower during squatting (p<0.033). Post-operative joint stiffness was greater in extension than in flexion, without any correlation with surgical competence. Although trained at different institutions, all fellows followed certain standard intraoperative guidelines during TKA, such as achieving neutral tibial abduction and avoiding internal tibial rotation,. 4. albeit at a static knee flexion angle. However, post-operative joint kinematics for dynamic motions revealed a strong correlation with surgical competence, i.e. kinematic variability over the range of passive flexion post-TKA was lower for more skilful surgeons. Moreover, actively loaded motions exhibited stark differences in post-operative kinematics as compared to those observed in passive motions. In vitro testing on the knee simulator also introduced the fellows to new quantitative parameters for post-operative joint assessment. In conclusion, the inclusion of cadaveric simulators replicating functional joint motions could help quantify training paradigms, thereby enhancing traditional orthopaedic training, as was also the unanimous opinion of all participating fellows in their positive feedback

For evaluating the impact of knee surgery, cadaveric knee simulators are commonly applied. However, most of the knee simulators are based on the Oxford type as originally described by Zavatsky (Zavatsky, J. of Biomechanics, 1997). These simulators mainly focus on the squatting motion. Although a wide range of flexion angles can be examined while performing this motion, the significance for activities of daily living is limited. To that extent a new knee simulator has recently been developed at Ghent University. In this simulator, the ankle motion is dynamically controlled in the sagittal plane; both in the proximal/distal direction and the anterior/posterior direction. As a result, this simulator allows simulating random motion patterns, e.g. cycling, stair ascent and descent, … The ankle translation is unrestrained in the medial/lateral direction. In addition, all rotational degrees of freedom are unrestrained at the ankle, resulting in four degrees of freedom at the ankle. The hip adds one rotational degree of freedom being the rotation in the sagittal plane. This leaves 5 degrees of freedom (DOF) to the knee; the sixth being flexion/extension that is controlled by the actuators at the ankle. During the simulation of different motion patterns, the quadriceps and hamstring force are actively controlled to mimic realistic conditions obtained through musculoskeletal simulations. In this study, five cadaveric experiments have been performed on the simulator. While mounting the cadaveric specimens in the test rig, the initial alignment remains crucial. Whilst the rig leaves 5 DOF to the knee, it is important to restore the anatomical position of the hip and ankle. To minimize the impact of the mounting procedure, cadaver specific 3D printed guides are used to assure the alignment of the cadaver in the test rig. As a result, the kinematics are more likely to represent physiological conditions. These kinematics have been evaluated in accordance to the methodology described by Grood&Suntay (Grood & Suntay, Transactions of the ASME, 1983). Therefore, a CT scan of the examined knee is combined with motion tracking data from rigidly attached markers on both the femur and the tibia. The cadaveric knees have been subjected to a variety of motion patterns, i.e. squatting and cycling. The squatting experiments provide evidence that the knee simulator creates adequate boundary conditions as the kinematic patterns coincide with literature reportings. The cycling experiments however significantly differ from the squatting patterns. Most noteworthy is the difference in terms of internal/external rotation for these native knees (Figure 1). This internal/external rotations is highly fluctuating from flexion to extension. This is understood as the quadriceps force is not constant during the extension phase, representing physiological conditions. Conclusion. Significant difference in knee kinematics between squatting and cycling indicates the importance of testing a variety of conditions. Furthermore, this reveals the need to study clinically relevant motion patterns, selected from patient reported outcomes