There are several important findings in this study. In the thin group, the MB was located more posteriorly at 90º flexion, the tibia was less internally rotated at 90º flexion, and the knee was slightly varus during almost the almost entire range of knee flexion compared with such measurements in the thick group. However, there were no significant differences in the MB movement and the tibial internal rotation angle at maximum knee extension and flexion.
A functional normal knee has a medial pivot motion and a bicondylar rollback motion during knee flexion [19, 20], and this combination enables the knee to move comfortably and flex deeply. This study showed that the medial contact point moved posteriorly, particularly after 90º flexion, and the medial posterior movement was recognized as a bicondylar rollback movement in the entire knee kinematics. In this study, the medial MB in the thin group was located more posteriorly at 90º flexion than that in the thick group, and the tibial internal rotational angle in the thin group was smaller at 90º flexion. Therefore, if the lateral contact point similarly moved posteriorly in both groups, these results could be interpreted as a bicondylar rollback occurring earlier in the thin group and a larger medial pivot motion until 90º in the thick group. In total knee arthroplasty, medial knee stability in the mid flexion angle has been reported as an important factor resulting in better postoperative clinical outcomes, and medial pivot motion in mid flexion is also reported to be essential to successful total knee arthroplasty [12, 21]. However, there is no evidence on medial pivot motion and postoperative clinical outcomes for UKA. Additionally, there were no significant differences in the posterior MB movement and the tibial internal rotation at maximum knee flexion between the thick and thin MBs that differed by 1 mm in thickness. Therefore, the tibia in the thin group rotated internally at a deep knee flexion angle. These kinematic differences also might influence postoperative clinical results and bearing dislocation. However, in this study, we were unable to compare the postoperative clinical outcomes between the two groups because the two groups only existed intraoperatively, and we adopted the final MB from both groups. Therefore, further investigation is necessary to reveal the entire knee kinematics situation and the relationship between intraoperative kinematics and postoperative clinical outcomes in UKA. Such future studies might help determine the ideal bearing thickness when choosing between a 1-mm difference.
Postoperative bearing dislocation is one of the main reasons for revision surgery after Oxford UKA [22]; however, the precise mechanisms causing this condition remain unknown. Bae and Lewold et al. mentioned that bearing dislocation could be attributed to component malposition and soft tissue imbalance with subsequent maltracking of the meniscal bearing [23, 24]. However, Lewold did not mention what maltracking of the MB indicated in their reports [24] and Bae et al. assumed that MB posterior overhang from the posterior edge of the tibial component could induce bearing dislocation [23]. Jamshed et al. reported a 180º bearing spin motion before the posterior bearing dislocation, and they mentioned that potential bearing spin motion could occur before a bearing dislocation [25]. Therefore, the intraoperative bearing movement is important. Kawaguchi et al. reported that the component position influenced the intraoperative MB movement, and they mentioned that MBs whose femoral components were set laterally tended to move posteriorly while in contact with the lateral wall [10]. MBs that are located beside the lateral wall did not tend to spin out; therefore, the component position could be an important factor for not only the intraoperative MB movement but also the bearing dislocation. Conversely, in this study, there was no significant difference in the distance between the MB tibial lateral wall or in the bearing rotation between the two groups. Therefore, the 1-mm difference in bearing thickness did not influence the relationship between the MB and the tibial lateral wall or the bearing rotation during passive knee flexion. However, the spin out stress test and the rollover sleep test (ROS test) [18] were performed to confirm the tendency of the bearing dislocation in this study before reaching a final decision on the bearing thickness. In the spin out stress test, the bearing was manually forced to rotate internally if the bearing had a tendency to rotate over 90º. Additionally, in the ROS test, the knee was forced into the valgus position, and the femur applied stress on the medial aspect of the tibial bearing, causing elevation of the lateral edge of the bearing to confirm whether a bearing has a tendency to dislocate into the intercondylar ridge. In these procedures, there were some unacceptable cases in which bearing dislocation occurred easily in the thin group; thus, the thicker bearing was chosen as the final bearing, as shown in Table 6. In future studies, the MB movement and the knee kinematics should be evaluated in not only passive knee flexion but also in these dislocation confirming tests.
When assessing coronal alignment in Oxford UKAs, a valgus correction should be performed carefully because overcorrected coronal valgus alignment could induce the progression of arthritis on the lateral side [26], and lateral osteoarthritis progression was one of the primary reasons for revision surgery [8]. In this study, the thick group displayed a greater valgus knee angle at each knee flexion angle except for 30º and 45º; however, the difference in the valgus knee angle was an average of approximately one degree. Misir et al. revealed a difference of approximately 3.6º in the postoperative tibiofemoral angle after Oxford UKA between the lateral OA progressed group and the non-progressed group [27]; thus, the difference in this angle between the two groups was much smaller in this study than the difference in their study. Additionally, Ro et al. compared complications after Oxford Phase III UKA between Asian and Western patients and reported that although the total reoperation rates did not differ between the two populations, reoperation for bearing dislocation was more likely to occur in Asian patients than in Western patients whereas reoperation for lateral knee OA progression was more likely to occur in Western patients than in Asian patients [5]. However, overcorrection of coronal alignment after Oxford UKA should not be neglected in Asian patients. Even if the surgeon cannot determine whether to choose the thin or thick MB that differ by 1 mm, this study could give surgeons the information that the difference influence one degree in coronal alignment, even in Asian patients with varus knee deformities and this information could aid the surgeons to decide the bearing thickness.
This study has some limitations. The first limitation is that intraoperative bearing movements were evaluated with a trial MB, which differs slightly from an actual MB. An actual MB is an ‘anatomic’ bearing with an extended lateral edge. However, actual tibial components do not have a scale on the surface, and MB movement cannot be evaluated with actual MBs and tibial components. Second, there could be an implantation error between the trial components and actual components. Implantation errors were checked with an intraoperative navigation system, with its alignment adjusted as little as possible. Third, we never experienced a postoperative bearing dislocation in this series, so the reasons for bearing dislocations remain unknown. Fourth, we did not distinguish between osteoarthritis and osteonecrosis. Further research with a larger sample should be conducted in the future.