Well-positioned lateral unicompartmental arthroplasty
Given the relative prevalence of isolated medial versus lateral compartment osteoarthritis, medial UKA is much more common than lateral UKA [8, 16, 17]. Overall, the most common causes for medial UKA failure are aseptic loosening (36%) and progression of osteoarthritis in the lateral compartment (20%), while the most common causes for lateral UKA failure are progression of osteoarthritis (29%), aseptic loosening (23%), and bearing dislocation (10%) [7, 18]. Despite the discrepancy in the prevalence of medial vs. lateral UKA, there are no significant differences in survival at 5, 10, or 15 years . Many studies analyze both medial and lateral UKA together, but the anatomic and biomechanical differences between compartments contribute to different failure modes, particularly in regard to mobile-bearing dislocations [7, 8, 19–23]. Differences including the convexity of the lateral condyle, the relative laxity of the lateral collateral ligament, and the medial pivot of the knee during flexion make mobile-bearing implants more prone to dislocation in lateral UKA, especially in the early postoperative timeframe [7, 19–25]. For this reason, some authors recommend using only fixed-bearing implants when considering a lateral UKA [24–26].
Overall, there is no difference in reoperation rates between fixed and mobile-bearing UKA [27, 28]. In a meta-analysis by Ko et al., progression of osteoarthritis and aseptic loosening were the most common causes of reoperation in mobile-bearing UKA, while polyethylene wear was the predominant cause of reoperation in the fixed-bearing group. The overall time to reoperation was shorter for mobile bearings, compared to fixed bearings, but the timing of this between groups depended on cause. Aseptic loosening and tibial component subsidence caused earlier revision in the fixed-bearing group. Progression of arthritis and unexplained pain led to similar timing of reoperation between mobile-bearing and fixed-bearing groups. Progression of arthritis may be more common in mobile-bearing groups due to surgeons’ desire to avoid mobile-bearing dislocation by putting in a “tight” knee and risk a slight overcorrection in alignment . In contrast, fixed-bearing implants may better tolerate under-correction to offload the other compartment .
Conversion rates for both lateral and medial UKA are low, with an approximately 1% annual revision rate . Conversion procedures can be technically demanding and often require the use of bone graft or augmentation to supplement bony deficits [31–34]. Most studies have found that the posterior cruciate ligament (PCL) can still be retained in converting to a total knee arthroplasty (TKA) [31–34]. Outcomes for revision of failed UKA have been shown to be equivalent if not superior to revision of failed primary TKA and are similar in long-term outcomes to primary TKAs [34, 35].
Modes of Failure
Progression of Arthritis
Progression of osteoarthritis in the remaining compartments is one of the most common modes of failure. Multiple studies cite the rate of UKA failure due to progression of arthritis between 1% and 9% [3, 36]. However, among UKA failures, progression of arthritis accounts for 15–50% of failures and is the most common mid- to late-term mode of failure [3, 5–7]. Patient selection and failure to follow specific indications may contribute, as patients with inflammatory arthritis, higher American Society of Anesthesiologists (ASA) score, and obesity are at a higher risk of developing adjacent compartment disease [37–39].
The clinical significance of patellofemoral joint degeneration is controversial. While it was initially thought that patellofemoral arthritis could contribute to anterior knee pain and patient dissatisfaction, recent literature has shown no difference in functional outcomes or reoperation rates in patients with patellofemoral arthritis documented at the time of surgery . An oversized femoral component can also lead to patellofemoral impingement, which may be more symptomatic than patellofemoral arthritis . This is more common with lateral UKA . Interestingly, recent studies have demonstrated improved patellofemoral joint congruence following UKA . The recommended management of symptomatic adjacent compartment degeneration is revision to total knee arthroplasty .
Polyethylene wear is a less common mode of UKA failure but still accounts for 12–25% of UKA failures . It usually presents as a late mode of failure, after 8 years, but early cases of catastrophic failure have also been reported [47, 48]. Technical factors that contribute to polyethylene wear include component malposition and under-correction of deformity [13, 47, 49]. Implant-specific factors associated with polyethylene wear include thickness less than 6 mm, lack of design conformity, and flaws in the manufacturing and sterilization of PE [47, 49]. Debris from polyethylene causes an osteolytic reaction at the bone-implant interface and can affect alignment and stability of the implant. Uneven load distribution from component malalignment/instability can accelerate aseptic loosening . Increased ligamentous laxity and subluxation of the tibiofemoral implant surface can also contribute to increased polyethylene wear . In addition, progression of osteoarthritis may also cause attenuation of the anterior cruciate ligament, leading to increased subluxation as well . Subluxation of the implants concentrates force over the peripheral aspect of the tibial component, which is also where the polyethylene layer is the thinnest . Improvements in wear-characteristics of polyethylene may decrease the prevalence of UKA failure from polyethylene wear. Management options of polyethylene wear include polyethylene exchange or revision to total knee arthroplasty.
Mobile-bearing dislocation accounts for 1.5–4.6% of UKA failures . These dislocations more commonly affect the lateral compartment secondary to the increased laxity of the lateral collateral ligament (LCL), the convexity of the lateral tibial condyle, and the medial femoral rollback during flexion [5, 7, 19–22, 49]. Medial mobile-bearing dislocations can occur in the setting of unbalanced flexion/extension gaps, instability due to medial collateral ligament (MCL) injury, or component malposition with impingement of the insert on the adjacent bone [46, 50]. In contrast to total knee arthroplasty, soft tissue releases are not recommended in UKA due to risk of instability and bearing dislocation, as the goal of UKA is to restore ligament tension to normal, thereby restoring knee kinematics . Joint instability following UKA can also contribute to early polyethylene wear and aseptic loosening, which can, in turn, increase the risk of mobile-bearing dislocation . Management for bearing dislocation includes revision UKA with a fixed bearing or conversion to total knee arthroplasty .
Tibial subsidence is the most common cause of periprosthetic fracture following UKA and accounts for 3.6–10.4% of UKA failures [6, 51]. Tibial collapse commonly presents as a late complication and is more common in elderly patients, suggesting osteoporosis as a contributing factor [51, 52]. It most commonly affects the medial tibial plateau due to increased pressure and load . Technical factors, which may contribute to tibial subsidence , include excessive posterior tibial slope [52, 53]. Surface area of the tibial component and depth of the tibial resection may also play a role, but this has not been demonstrated in the literature . Tibial collapse is also more commonly associated with fixed-bearing all-polyethylene implants. These implants have higher contact stresses at localized points in the anterior and medial tibia, leading to localized edge loading and tibial collapse . Management options for tibial collapse include percutaneous screw fixation and revision to TKA and may require cement, augments, cones, and stems depending on the degree of bone loss, status of the other knee compartments, and degree of deformity [31–34, 46, 52, 55].
Infection following UKA occurs at a rate of 0.2–1% . While this rate is slightly lower than that of total knee arthroplasty, UKA infections are unique in that they involve both the prosthesis and native cartilage . The diagnostic workup of infection in UKA is based on that of total knee arthroplasty with preoperative laboratory studies used to guide potential aspiration [57, 58]. Given the involvement of both native cartilage and prosthesis, slightly different thresholds for the diagnosis of infection have been proposed. In a study of 259 patients undergoing revision UKA, Schwartz et al.  found a 10.8% infection rate and proposed cutoff values of the following: ESR: 27 mm/h, CRP: 14 mg/L, synovial fluid WBC count of 6200 cells/μL, and 60% neutrophils . The proposed cutoff for aspirate leukocyte count is higher than that of total knee arthroplasty. This may be attributable to the involvement of more of the native cartilage but requires further investigation and validation . The causative organisms in UKA infections are similar to those in total knee arthroplasty infection with coagulase-negative Staphylococcus, S. aureus, group B Streptococcus, E. coli, and P. acnes among the most common organisms [56, 59]. Management of UKA infection is similar to that of total knee arthroplasty. Acute infections can be managed with irrigation, debridement , polyethylene liner exchange, and antibiotics, while chronic infections require irrigation and debridement with antibiotic spacer, antibiotics, and conversion to total knee arthroplasty . Labruyere et al. proposed synovectomy with one-stage conversion to total knee arthroplasty and 3 months of antibiotics as a reasonable alternative to two-stage conversion to total knee arthroplasty, with no infection recurrence and good functional outcomes at early follow-up .
Unexplained pain is an important cause for revision after UKA. While etiology is unknown, it accounts for up to 23% of revision surgery according to registry data from England and Wales . This is significantly higher than the rates of revision for total knee arthroplasty for unexplained pain, which is estimated to be about 9% . Overall, the rates of the unexplained pain following UKA vary from 5% to 15% [5–7]. While the etiologies may differ on a patient-by-patient basis, proposed etiologies include loose bodies, cement extrusion, meniscal tears in the native compartment, and chronic regional pain syndrome or reflex sympathetic dystrophy . Due to the limited exposure and single-piece implants used in some systems, it can be technically challenging to remove extruded cement from the posterior aspect of a UKA. Given the hybrid of native cartilage and prosthesis, failure to restore normal joint alignment and mechanics can contribute to pain generators such as meniscal tears and loose bodies. Furthermore, all-polyethylene tibial components have been associated with higher rates of unexplained pain . This may be related to the higher rates of tibial collapse seen with fixed-bearing all-polyethylene tibial component due to higher load transfer to the tibia resulting in persistent bone remodeling . The threshold for revision due to unexplained pain may also differ based on surgeon experience and familiarity with UKA . In an examination of the New Zealand Joints Registry, Goodfellow et al. showed that in knees that had very poor outcome (Oxford Knee Score <20), 63% of UKAs went onto revision surgery, while only 12% of TKAs were revised . Management of the unexplained pain following UKA is surgeon-specific, but the conversion of a UKA to a total knee arthroplasty is less technically demanding than revision of a total knee arthroplasty. Thus, surgeons may have a lower threshold to offer revision to total knee arthroplasty as an option for patients with unclear pain generators following UKA [60, 62].
Patellofemoral arthroplasty (PFA) has seen an increase in popularity owing in large part to the development of second-generation PFA design and technique . First generation of complete PFA (replacement of both patella and trochlea) was performed using the inlay technique – replacing only the trochlear cartilage and leaving subchondral bone intact. As a result, position of the trochlear component was dictated based on the anatomy of the native trochlea and free-hand technique proved technically challenging [64–66]. While early 1- to 2-year outcomes were encouraging, first-generation PFA showed high rates of excessive wear in the trochlear groove and patellar maltracking [65–68]. Long-term follow-up, ranging from 5 to 20 years, showed revision rates of 25–60%, with failure secondary to patella maltracking, trochlear wear, and progression of tibiofemoral joint arthritis [67–70].
Second-generation PFA was developed as a result of the high percentage of failures seen with the inlay technique. With the second-generation onlay design, anterior femoral cuts are made and the trochlear prosthesis is seated within the anterior compartment of the knee. In addition, modifications to the trochlear implant design and shape allowed for decreased catching, better tracking angle, and congruity throughout range of motion . As a result of these changes , as well as careful patient selection, revision rates with second-generation techniques have been reported to be as low as 3–10% at 5- to 10-year follow-up [72–74]. With the onlay technique, revision due to patellar maltracking has been shown to be approximately 1–2% [73, 74]. The most common modes of failure for modern PFA are progression of tibiofemoral arthritis (38%), persistent anterior knee pain (16%), aseptic loosening (14%), and patellar maltracking (10%). Persistent anterior knee pain is the most common complaint among those with an early need for revision, while progression of tibiofemoral arthritis is the most common late mode of failure [64, 73, 75]. When comparing second-generation PFA and TKR, some studies have shown no difference in revision rates or pain while also reporting quicker recovery and higher activity scores in patients who underwent PFA [75, 76]. Treatment for failure of PFA is conversion to TKA, with multiple studies showing equivalent outcomes to primary TKA [77, 78].
UKA has developed over the years into a successful and predictable procedure, with survival rates greater than 90% at 10, 15, and 20 years. Aseptic loosening and adjacent compartment arthritis account for the majority of UKA revisions. Mobile-bearing dislocation, tibial collapse, and infection account for the remaining implant failures. Patellofemoral arthroplasty is most often revised due to persistent anterior knee pain or progression of tibiofemoral arthritis. Unexplained pain following UKA is another common cause of revision but may be biased by the relative ease of converting a UKA to a total knee arthroplasty. Conversion to total knee arthroplasty is generally recommended for implant failure; however, less invasive management such as irrigation and debridement with polyethylene liner exchange may be appropriate in select circumstances such as acute infection or polyethylene wear.