Discussion
In this systematic review, the most common failure modes in lateral UKA review were OA progression (29%), aseptic loosening (23%), and bearing dislocation (10%). Progression of OA and bearing dislocation were the most common modes of failure in cohort studies (36% and 17%, respectively), while aseptic loosening and OA progression were the most common failure modes in registry-based studies (28% and 24%, respectively).
As mentioned above, there are differences in anatomy and kinematics between the medial and lateral compartment. When the lateral UKA failure modes are compared with studies reporting medial UKA failure modes, differences in failure modes are seen.34 Siddiqui and Ahmad35 performed a systematic review of outcomes after UKA revision and presented a table with the failure modes of included studies. Unfortunately they did not report the ratio of medial and lateral UKA. However, when assuming an average percentage of 90% to 95% of medial UKA,6,7,36 the main failure mode in their review in 17 out of 21 studies was aseptic loosening. Indeed, a recent systematic review on medial UKA failure modes showed that aseptic loosening is the most common cause of failure following this procedure.34 Similarly, a search through registry-based studies6,7 and large cohort studies37-40 that only reported medial UKA failures showed that the majority of these studies7,37-39 also reported aseptic loosening as the main cause of failure in medial UKA. When comparing the results of our systematic review of lateral UKA failures with the results of these studies of medial UKA failures, it seems that OA progression seems to play a more dominant role in failures of lateral UKA, while aseptic loosening seems to be more common in medial UKA.
Differences in anatomy and kinematics of the medial and lateral compartment can explain this. Malalignment of the joint is an important factor in the etiology of OA41,42 and biomechanical studies showed that this malalignment can cause decreased viability and further degenerative changes of cartilage of the knee.43 Hernigou and Deschamps44 showed that the alignment of the knee after medial UKA is an important factor in postoperative joint changes. They found that overcorrection of varus deformity during medial UKA surgery, measured by the hip-knee-ankle (HKA) angle, was associated with increased OA at the lateral condyle and less tibial wear of the medial UKA. Undercorrection of the varus caused an increase in tibial wear of polyethylene. Chatellard and colleagues45 found the same results in the correction of varus, measured by HKA. In addition, they found that when the prosthetic (medial) joint space was smaller than healthy (lateral) joint space, this was correlated with lower prosthesis survival. A smaller joint space at the healthy side was correlated with OA progression at the lateral compartment and tibial component wear.
These studies explain the mechanism of progression of OA and aseptic loosening. Harrington46 assessed the load in patients with valgus and varus deformity. Patients with a valgus deformity have high mechanical load on the lateral condyle during the static phase, but during the dynamic phase, a major part of this load shifts to the medial condyle. In the patients with varus deformity, the mechanical load was noted on the medial condyle during both the static and dynamic phase. Ohdera and colleagues47 advised, based on this biomechanical study and their own experiences, to correct the knee during lateral UKA to a slight valgus angle (5°-7°) to prevent OA progression at the medial side. van der List and colleagues48 similarly showed that undercorrection of 3° to 7° was correlated with better functional outcomes when compared to more neutral alignment. Moreover, Khamaisy and colleagues49 recently showed that overcorrection during UKA surgery is more common in lateral than medial UKA.
These studies are important to understanding why OA progression is more common as a failure mode in lateral UKA. The shift of mechanical load from the lateral to medial epicondyle during the dynamic phase also could explain why aseptic loosening is less common in lateral UKA. As Hernigou and Deschamps44 and Chatellard and colleagues45 stated, undercorrection of varus deformity in medial UKA is associated with higher mechanical load on the medial prosthesis side and smaller joint space width. These factors are correlated with mechanical failure of medial UKA. We think this process can be applied to lateral UKA, with the addition that the mechanical load is higher on the healthy medial compartment during the dynamic phase. This causes more forces on the healthy (medial) side in lateral UKA, and in medial UKA more forces on the prosthesis (medial) side, which results in more OA progression in lateral UKA and more aseptic loosening in medial UKA. This finding is consistent with the results of our review of more OA progression and less aseptic loosening in lateral UKA. This study also suggests that medial and lateral UKA should not be reported together in studies that present survivorship, failure modes, or clinical outcomes.
A large discrepancy was seen in bearing dislocation between cohort studies (17%) and registry-based studies (5%). When we take a closer look to the bearing dislocation failures in the cohort studies, most of the failures were reported in only 2 cohort studies.50,51 In a study by Pandit and colleagues,50 3 different prosthesis designs were used in 3 different time periods. In the first series of lateral UKA (1983-1991), 6 out of 51 (12%) bearings dislocated. In the second series (1998-2004), a modified technique was used and 3 out of 65 (5%) bearings dislocated. In the third series (2004-2008), a modified technique and a domed tibial component was used and only 1 out of 68 bearings dislocated (1%). In a study published in 1996, Gunther and colleagues51 also used surgical techniques and implants that were modified over the course of the study period. Because of these modified techniques, different implant designs, and year of publication, bearing dislocation most likely plays a smaller role than the 17% reported in the cohort studies. This discrepancy is a good example of the important role for the registries and registry-based studies in reporting failure modes and survivorship, especially in lateral UKA due to the low surgical frequency. Pabinger and colleagues52 recently performed a systematic review of cohort studies and registry-based studies in which they stated that the reliability in non-registry-based studies should be questioned and they considered registry-based studies superior in reporting UKA outcomes and revision rates. Furthermore, given the differences in anatomic and kinematic differences between the medial and lateral compartment and different failure modes between medial and lateral UKA, it would be better if future studies presented the medial and lateral failures separately. As stated above, most large cohort studies and especially annual registries currently do not report modes of failure of medial and lateral UKA separately.3,4,18-20
There are limitations in this study. First, this systematic review is not a full meta-analysis but a pooled analysis of collected study series and retrospective studies. Therefore, we cannot exclude sampling bias, confounders, and selection bias from the literature. We included all studies reporting failure modes of lateral UKA and excluded all case reports. We made a conscious choice about including all lateral UKA failures because this is the first systematic review of lateral UKA failure modes. Another limitation is that the follow-up period of the studies differed (Table 1) and we did not correct for the follow-up period. As stated in the example of bearing dislocations, some of these studies reported old or different techniques, while other, more recently published studies used more modified techniques11,29,53-56 Unfortunately, most studies did not report the time of arthroplasty survival and therefore we could not correct for the follow-up period.
In conclusion, progression of OA is the most common failure mode in lateral UKA, followed by aseptic loosening. Anatomic and kinematic factors such as alignment, mechanical forces during dynamic phase, and correction of valgus seem to play important roles in failure modes of lateral UKA. In the future, failure modes of medial and lateral UKA should be reported separately.
Am J Orthop. 2016;45(7):432-438, 462. Copyright Frontline Medical Communications Inc. 2016. All rights reserved.