Introduction

Unicompartmental knee arthroplasty (UKA) comprises 8–10% of all knee arthroplasties according to national registries [1,2,3,4]. Registries and cohort studies demonstrated lower survivorship of UKA compared to total knee arthroplasty (TKA) [5]. However, studies showed several advantages of UKA over TKA; lower infection and mortality rates, faster recovery, larger range of motion (ROM) and better patient-reported outcome measures (PROMs) [6,7,8,9]. Due to these advantages, as well as improvements in implant design and surgical techniques, and a better understanding of patient indications, UKA is gaining popularity [2,3,4]. The majority of all UKA procedures (85–95%) concern the medial compartment, while the surgical volume of the lateral UKA is limited [2, 4, 10]. Lateral UKA is performed less frequently as it is considered a more technically demanding surgical procedure and the prevalence of isolated lateral Osteoarthritis (OA) is lower compared to medial OA [11, 12].

The lateral tibiofemoral compartment has different anatomy and kinematics compared to the medial compartment, resulting in a greater anteroposterior translation during knee flexion. Furthermore, increased laxity in flexion is noted laterally [13,14,15]. Therefore, it has been suggested that implant design affects outcomes of lateral UKA to a greater extent than it may in medial UKA [16]. Currently, there are two bearing types available for UKA, mobile and fixed-bearing designs. Mobile-bearing designs have the theoretical advantage in restoring the biomechanics of the lateral compartment by allowing anteroposterior translation of the insert, which results in lower contact stresses on the polyethylene [14, 17]. However, due to the increased laxity, bearing dislocation has been reported frequently following lateral UKA surgery with mobile-bearing implants [16]. Therefore, in the setting of lateral UKA, the choice of bearing design remains controversial.

To our knowledge, limited evidence is available comparing outcomes of fixed and mobile-bearing lateral UKAs. Therefore, the primary aim of this study was to systematically assess survivorship of fixed and mobile-bearing lateral UKAs. Secondary, functional outcomes of both implant designs were evaluated. The hypothesis of this study was that fixed-bearing designs would show higher survivorship compared to mobile-bearing designs. Furthermore, it was expected that better functional outcomes would be reported following mobile-bearing lateral UKA.

Materials and methods

Search strategy

This systematic review was conducted according to PRISMA guidelines [18]. A comprehensive search was performed in Medline, EMBASE and Cochrane Library on February 12, 2018. Search terms included “unicompartmental”, “knee”, “arthroplasty” and the associated synonyms (“Appendix”). Additionally, common medical abbreviations and “lateral” were added. No limits on publication date or patient age were used. Search strategies are shown in “Appendix”. After removing duplicates, title and abstract of the studies were screened by two authors independently (JAB and LJK), considering the eligibility criteria (Table 1). References of included studies were checked for any missing studies. The third author (INS) was consulted in case of disagreement. Consensus was achieved with regard to inclusion or exclusion for all reviewed articles.

Table 1 Inclusion and exclusion criteria
Full size table

Data extraction

Data was extracted and collected in Microsoft Excel 2017 by one of the authors (JAB), and subsequently, checked for accuracy by another author (INS). The data from the included studies was divided into two groups based on bearing type: (1) fixed-bearing and (2) mobile-bearing lateral UKA. The first author, study design (retro- or prospective), study characteristics (publication year, country, time of inclusion, number of patients and knees), patient characteristics (age, sex, body mass index; BMI), arthroplasty (implant design, surgical approach, indication), implant survival data (revision, failure mode, follow-up) and functional outcomes (e.g., ROM, PROMs) were extracted. Authors were contacted for additional information when indication of the lateral UKA were unspecified. Results reported as medians were transformed to means by the method of Hozo et al. [19].

Implant survival data was transformed into annual revision rate to correct for different follow-up intervals between populations. This metric is defined as revision rate per observed component-year [5, 6]. Revision was defined as ‘a new operation in a previously resurfaced knee during which one or more of the components are exchanged, removed, or added’ according to the Swedish arthroplasty registry [4]. Therefore, additional medial UKA and patellofemoral arthroplasty for OA progression were both considered as a revision. Moreover, re-operations that were not clearly described were considered a revision.

Methodological quality assessment

Methodological quality assessment was performed independently by two authors (JAB and LJK) using the validated MINORS criteria (methodological index for non-randomized studies) [20]. The third author (INS) was consulted in case of disagreement.

Statistical analysis

Poisson-normal models with random effects were used to estimate pooled annual revision rate data separately for fixed and mobile-bearing lateral UKA cohorts. The log incidence rates of each cohort enabled the calculation of overall log incidence rates per bearing type. Pooled annual revision rates with 95% confidence intervals (CI) were obtained by back-transforming the log incidence rates. Analyses were conducted using R version 3.2.3 (R Foundation for Statistical Computing, Vienna, Austria) with Metafor version 1.9-8 (Maastricht University, Maastricht, Netherlands). In addition, a subgroup analysis was performed to determine the annual revision rate for the Oxford domed mobile-bearing design (Biomet UK ltd, Swindon, UK).

Primary lateral OA was the main indication in the majority of studies (> 75% of the cohort), however, in a proportion of the studies this information was lacking. A sensitivity analysis was conducted, as secondary OA could influence the result as it is associated with inferior outcomes [21, 22].

For pre- and postoperative functional outcomes, means and standard deviations (SD) for each bearing type were combined and reported as mean difference with 95% CI. These pooled analyses were performed with RevMan version 5.3 (Cochrane Reviews, London, UK) using the inverse variance method. If outcomes were reported with a range, the SD was calculated using the method by Walter and Yao [23]. A p value < 0.05 was considered to be statistically significant.

The heterogeneity between studies were determined by the χ2 and I2 statistic method in both the annual revision rate and functional outcome analyses. Heterogeneity was considered with a I2 value of 25% to be low; 50%, moderate; and 75%, high [5].

Results

A total of 28 studies were included, representing 2265 lateral UKAs (Fig. 1) in 19 fixed-bearing and nine mobile-bearing studies [11, 16, 17, 22, 24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47]. Seventy-eight percent of the mobile-bearing studies used the Oxford domed design, while a variety of fixed-bearing implants were used. Nineteen (70%) studies used a lateral parapatellar approach, and eight (30%) used another approach (e.g., medial parapatellar, traditional TKA) or included multiple approaches over the course of the study. There was female predominance (range 52–93%). Mean age ranged from 53 to 74 years and mean BMI from 25 to 33 kg/m2 (Table 2).

Fig. 1
figure 1

Preferred reporting items for systematic reviews and meta-analyses flow diagram. UKA unicompartmental knee arthroplasty, OA osteoarthritis

Full size image
Table 2 Study characteristics
Full size table

Quality of the studies

Twenty (71%) retrospective and eight (29%) prospective studies were included. The mean MINORS score of the comparative studies was 16.1 (range 13–18), and 9.4 (range 7–12) for the non-comparative studies, representing 67% and 59% of the maximum score, respectively (Table 2). None of the studies were blinded and only three (11%) reported prospective calculation of the sample size.

Annual revision rates

The overall annual revision rate of the fixed-bearing group was 0.94 (95% CI 0.66–1.33) and 2.16 (95% CI 1.54–3.04) for the mobile-bearing group (Table 3; Figs. 2, 3). A subgroup analysis was performed for the domed mobile-bearing design; showing an overall annual revision rate of 1.81 (Table 3; Fig. 4). Annual revision rates of each study were converted to survival rates and plotted (Fig. 5). Overall, heterogeneity was low among the fixed-bearing studies and moderate for mobile-bearing studies (Table 3). The sensitivity analysis for studies with unspecified indications for lateral UKA showed no differences in survival compared to studies that did specify their indication. The distribution of modes of failure is shown in Table 4. A total of seven fixed-bearing and five mobile-bearing studies specified the type of revisions performed [17, 24,25,26, 29, 31, 32, 35, 37, 38, 41, 42, 46]. It was noted that fixed-bearing lateral UKAs were more frequently converted to TKA (67.4% of mobile-bearing versus 41.9% of mobile-bearing), while tibial bearing or insert exchange was more commonly reported after mobile-bearing lateral UKA (29% of mobile-bearing versus 0% of fixed-bearing).

Table 3 Annual revision rate per bearing type or design
Full size table
Fig. 2
figure 2

Annual revision rates of fixed-bearing studies. CI confidence interval

Full size image
Fig. 3
figure 3

Annual revision rates of mobile-bearing studies. CI confidence interval

Full size image
Fig. 4
figure 4

Annual revision rates of domed mobile-bearing cohorts. CI confidence interval

Full size image
Fig. 5
figure 5

Survival rates with length of follow-up from all included studies per bearing type. Diameter of circle representing cohort size

Full size image
Table 4 Failure mode per bearing type or design
Full size table

Functional outcomes

A total of 11 studies reported pre- and postoperative PROMs and 13 studies reported ROM [11, 16, 17, 24, 26,27,28,29, 32, 35, 37, 38, 40, 42, 43, 45]. Overall, no statistically significant differences were found in KSS (Knee Society Score) knee, KSS function, KSS total, OKS (Oxford Knee Scores), nor in ROM between the two designs. Overall, heterogeneity was high among the fixed-bearing studies and high to low for mobile-bearing studies (Table 5).

Table 5 Effect of lateral unicompartmental knee arthroplasty on different functional outcomes
Full size table

Discussion

The main finding of this study was that the risk of revision is lower after lateral UKA with fixed-bearing designs than mobile-bearing designs. The annual revision rate of the mobile-bearing was 2.16, domed mobile-bearing 1.81 and fixed-bearing designs 0.94, corresponding to extrapolated 5-year survival rates of 89%, 91% and 95%, respectively. This dissimilarity between the two bearing types could be a consequence of a higher proportion of dislocations after mobile-bearing lateral UKAs (44%) compared to fixed-bearing lateral UKAs, although the new domed mobile-bearing design reduces the number of bearing dislocations. Progression of OA in the medial compartment was observed as mode of failure in both fixed-bearing and mobile-bearing designs (53% and 19%, respectively). Furthermore, good-to-excellent PROMs and ROM were noted with both bearing types, which did not support our hypothesis favoring the use of mobile-bearing designs based on functional outcome. This study highlights the importance of assessing survivorship and functional outcomes in lateral UKAs per bearing type using a systematic approach, due to the low prevalence of this procedure.

Currently, studies comparing fixed and mobile-bearing lateral UKA survivorship are limited. One large registry-based study by Baker et al., including 2052 lateral UKAs, found no statistical difference in survival between fixed and mobile-bearing designs [10]. Contrary to our definition for revision, the authors did not differentiate between individual components for each prosthesis, which could lead to smaller differences in survival rates between the two bearing types. Therefore, it is important to bear in mind that the definition of revision could influence study results. Our extrapolated 5-year survival rate of domed mobile-bearing was higher when compared to a recent Danish registry, which included 52 domed mobile-bearing implants (91% versus 87.4%, respectively) [48]. To our knowledge, no other registry or systematic review has assessed lateral UKA survivorship per bearing type. In summary, this overview stresses the need for studies and registries to assess survivorship of lateral UKAs per bearing type, as data are very limited.

Regarding modes of failure, bearing dislocation was frequently noted in mobile-bearing designs (44%), whereas progression of OA in the medial compartment was common in both fixed and mobile-bearing designs (53% and 19%, respectively). Most of the included studies failed to report the time of revision with corresponding mode of failure. Therefore, in this study, modes of failure per bearing type could not be corrected by follow-up period. This may explain the higher percentage of progression of medial OA in the fixed-bearing group, as their average follow-up is longer and progression of OA often occurs later after the initial surgery. The high percentage of bearing dislocations in mobile-bearing lateral UKAs may be due to larger joint distraction in flexion laterally compared to medially (7 mm versus 2 mm) [15]. To lower the rates of bearing dislocation, the domed mobile-bearing tibial implant has been introduced. This implant potentially reduces the incidence, as it requires more distraction before the polyethylene insert dislocates, however, it does not eliminate the possibility of bearing dislocation [13, 16]. Overall, the available literature implicates that the benefit of using fixed-bearing designs is that bearing dislocation cannot occur.

When reviewing mobile-bearing results, surprisingly no dislocations were observed in two studies. Liebs et al. used a mobile-bearing of which the insert slides into a groove from anterior to posterior, while medial–lateral translation is restrained. Consequently, dislocation did not occur; however, aseptic loosening was frequently noted, and therefore, this implant is no longer used [30, 33, 49]. The study by Van Duren et al. used a trans-patellar approach to optimize the access for the vertical cut of the tibia [28]. This approach allows the surgeon to place the tibial component and bearing in a potentially more optimal position. No dislocations of the domed mobile-bearing were reported at a relatively short median follow-up of 27.4 months. Several studies suggested that component alignment is critical to reduce the risk of dislocation in mobile-bearing designs [50, 51]. Gulati et al. recommended after radiographic evaluation of knees with dislocated bearings that overstuffing should be avoided and the femoral component needs to be neutrally aligned in flexion.

Several mobile-bearing studies have managed bearing dislocation successfully by replacing the bearing with a thicker one. In addition, the revision method by Weston-Simons et al. was used, in which the bearing was exchanged and two to three screws were inserted with their heads above the medial wall of the tibial plateau to prevent recurrent dislocation [17]. According to the Australian registry, these types of revision are classified as a minor revision, but have a higher risk of re-revision compared to revision to TKA [2, 52]. Therefore, UKA surgeons need to carefully consider if bearing exchange is a useful option when revising a UKA on the lateral side for bearing dislocation.

To prevent progression of OA in the medial compartment following mobile and fixed-bearing lateral UKAs, it is generally stated that overcorrection should be avoided [24, 35, 38, 43]. Ohdera et al. suggested a valgus aligned mechanical axis between 5°–7° should be aimed in lateral UKA surgery [38]. Furthermore, Van der List et al. showed that postoperative valgus of 3°–7° was correlated with better functional outcomes than more neutral aligned knees [53]. However, a cautious approach is needed since MUKA studies reported that undercorrection is associated with polyethylene wear [54, 55]. Based on the results of this study and those reported by Baker et al., polyethylene wear was less frequently reported as a mode of failure after lateral UKA [10]. Future studies are necessary to evaluate the association between degree of valgus and polyethylene wear following lateral UKAs.

In the present study, no statistically significant differences were found in functional outcomes between both bearing types. Mobile-bearing UKAs may have theoretical biomechanical advantages; however, this did not affect the functional outcomes after surgery. Only two small comparative studies assessed functional outcomes, and similarly, showed no statistically significant differences in OKS and ROM between both designs [30, 36].

This study has several limitations. Although the metric annual revision rate corrects for different follow-up intervals between studies, it relies on the assumption that the revision rate remains constant over time. Therefore, mobile-bearing findings have to be interpreted with more caution than those for fixed-bearing, because annual revision rate of mobile-bearing were only based on short- to mid-term results. Furthermore, despite the majority of studies having > 70% patients with primary OA and having performed a sensitivity analysis based on indication, considerable variability of the indication for each procedure existed. Another limitation is that a majority of the studies consisted of small cohorts with low to moderate quality. In addition, several cohort studies in the domed mobile-bearing group have led to concerns about the reliability due to developer bias, therefore, assessing survivorship and functional outcomes with registry data may be helpful. However, only one annual registry reported survivorship of the domed mobile-bearing design and one registry-based study reported results of both bearing types separately. Hence, this study provides insights to the current literature.

Conclusion

In conclusion, mobile-bearing lateral UKA have a higher rate of revision compared to fixed-bearing lateral UKA with regard to short- to mid-term survivorship; however, the clinical outcomes are similar. Despite the improvements in mobile-bearing implants with a domed shaped design, short- to mid-term survivorship remains inferior to the fixed-bearing designs due to a high percentage of bearing dislocations leading to revision. Other common failure modes in both the fixed and mobile-bearing designs were progression of OA. As a result of the moderate evidence included in this study, future registry data are needed to confirm these findings. Nonetheless, the results of this study suggest a preference of using fixed-bearing implants for isolated lateral knee OA.