The most important findings in this study on UKA were that the intraoperative tibial rotation during knee flexion was larger in nonelderly patients than in elderly patients, and the intraoperative tibial internal rotation had a positive influence on the postoperative clinical outcomes only in nonelderly patients.
Regarding the kinematics in Oxford UKA, Kono et al.  compared the in vivo preoperative and postoperative kinematics in Oxford UKA using 2D-3D fluoroscopic surveillance and reported that the tibiofemoral preoperative rotational kinematics was maintained postoperatively after the Oxford mobile-bearing UKA procedure. In addition, Oxford mobile-bearing UKA has been suggested to be superior to fixed-bearing UKA in restoring the normal tibiofemoral biomechanics [27-29]. However, no study has investigated the age difference in perioperative kinematics in UKA. To our knowledge, our study is the first to demonstrate the age difference in intraoperative tibiofemoral rotational kinematics during knee flexion, reporting that the tibial internal rotation was larger in nonelderly patients. In normal knees, the tibial internal rotation during knee flexion has been reported to be a “medial pivot motion,” and this tibial internal rotation could be generated by the four-bar linkage of both cruciate and collateral ligaments [30-32]. Therefore, it was believed that the difference in internal rotation between two groups could be the condition of ligaments and related soft tissues. However, the normal axial rotation decreased proportionately with the progression of medial-grade OA . Therefore, we compared the grade of OA between the elderly and nonelderly patients and observed that the preoperative bone varus deformity in the X-ray was not different between the two patient groups. In addition, this finding supported our theory that the difference in the tibial internal rotation was derived from the quality of related soft tissues.
Nevertheless, UKA has become a standard surgery for not only middle-aged patients but also elderly patients [4, 5, 34, 35]. Iacono et al.  evaluated the mid- to long-term clinical and radiographic results after Oxford mobile-bearing UKA in elderly patients (aged ≥75 years) and reported that their postoperative range of motion, Oxford Knee Score, Knee Society Score, and WOMAC score were satisfactory and the tibial component alignment was still maintained at the 9-year follow-up. They concluded that UKA was a viable option to treat medial OA even for elderly patients in terms of pain relief and activity scores. In contrast, Siman et al.  reported that in elderly patients (aged ≥75 years), UKA could not result in a superior Knee Society Score to that in patients undergoing TKA, although patients undergoing UKA had shorter hospital stay, shorter operative time, and lower estimated blood loss than patients undergoing TKA. Furthermore, Kennedy et al.  reported that the Oxford Knee Score after UKA in elderly patients was significantly lower than that in younger patients. Nevertheless, these studies did not distinguish clinical pain relief and clinical recovery of activity postoperatively, because elderly patients with lower muscle strength could have lower potential in recovery of activity than nonelderly patients who have higher muscle strength. The difference between TKA and UKA could disappear because of the lower activity recovery in elderly patients. Therefore, it is necessary to evaluate postoperative clinical results separately with the pain relief and activity components when evaluating elderly patients. From this viewpoint, the present study is meaningful and our results are reasonable, that is, pain relief and flexion angle after UKA were achieved in both elderly and nonelderly patients, and the activity scores after UKA were greater in nonelderly patients than in elderly patients. However, Fabre-Aubrespy et al.  evaluated the clinical outcomes of elderly patients (aged ≥75 years) who underwent fixed-bearing UKA and distinguished pain relief scores from activity scores. They reported that not only pain relief scores (Pain and Symptom in the KOOS, Knee Society Score) but also the recovery of activity scores (NKFS and ADL in KOOS) in their study were better for elderly patients treated with UKA than for elderly patients treated with TKA. Their results confirmed that the activity score after UKA in elderly patients was less than that in nonelderly patients, but more than that in elderly patients undergoing TKA, thereby indicating that UKA was a viable surgical option even for elderly patients in terms of pain relief and recovery of activity.
Importantly, several recent studies have emphasized the extent of intraoperative tibial rotation in TKA [14, 15, 36], and the majority of them have demonstrated a relationship between the tibial internal rotational angle and the postoperative flexion angle. Furthermore, the majority of studies have focused on the tibial internal rotational angle during the latter half of knee flexion [14, 15, 36]. In the present study, the intraoperative tibial internal rotation during the entire knee flexion correlated positively with the pain subscale in the KOOS, and the internal rotational angle from 90° to full flexion correlated positively with the pain and QOL subscales in the KOOS only in nonelderly patients. The primary reason for the correlation between internal rotation during the latter half of knee flexion and clinical outcomes is believed to be the medial pivot motion of UKA. Kono et al.  described the postoperative non-weight-bearing kinematics of UKA in 24 functionally well knees and reported the tibial internal rotation during the latter half of knee flexion. In their study, the lateral contact point constantly moved posteriorly, whereas the medial contact point moved only slightly, especially during the latter half of knee flexion. This would be the so-called medial pivot motion associated with Oxford UKA in a non-weight-bearing position. Non-weight-bearing kinematics of the knee after surgery is expected to be similar to intraoperative kinematics . We could not evaluate the accurate anterior–posterior position of the tibia relative to the tibia, in particular, for each component from our navigation data, and it was difficult to distinguish whether intraoperative tibial internal rotation was the central pivot motion or the medial pivot motion. Nevertheless, our results in nonelderly patients appear remarkably similar to previous reports that showed that the medial pivot motion during the operation correlated with better subjective scores in TKA . Therefore, we believed that in UKA for nonelderly patients, the intraoperative medial pivot motion during the latter half of knee flexion resulted in excellent postoperative KOOSs. However, we could not identify any correlations between the intraoperative tibial rotation and clinical results in elderly patients. These patients had smaller intraoperative tibial internal rotation than nonelderly patients, but their postoperative patient-reported outcomes, except for activity scales, were not different from those of nonelderly patients. This implied that the relationship between kinematics and clinical outcomes in elderly patients was weaker than that in nonelderly patients, and multiple factors, such as painful duration, muscle strength, and lumbar disorders, could influence postoperative patient-reported outcomes in elderly patients. Nonetheless, we could not find any correlation between intraoperative tibial internal rotation and postoperative knee flexion angle in both our study groups, unlike that in previous studies on TKA [14, 15, 36]. The reason for this difference between TKA and UKA is that the postoperative flexion angle after UKA was relatively good in almost all patients. Therefore, the difference in intraoperative rotation did not affect the postoperative knee flexion angle.
This study has certain limitations. First, the patients were divided into two groups based on the age of 75 years; therefore, all those aged <75 years were grouped into one group. This relatively young group contained patients aged 54–74 years, and differences in clinical outcomes and intraoperative kinematics would certainly exist within this group. Therefore, a further investigation is planned with a larger cohort. Second, the follow-up period was relatively short. However, we experienced only a few problems related to loosening, polyethylene wear, or breakage of Oxford UKA. Hence, a larger sample size with a longer follow-up period is required. Third, although the KOOS is a valid, reliable, and responsive outcome measure in patients with knee arthroplasty, minimal clinically important differences in the KOOS were not apparent. Fourth, the procedures were performed by five knee surgeons, which raises the possibility of interobserver bias. Fifth, we did not evaluate the alignment of the component. Sixth, preoperative KOOS data were not available for all patients; however, p values in the preoperative KOOS were relatively high.