Unicompartmental knee arthroplasty: state of the art

To assess whether or not a patient is indicated for UKA depends on many factors. On physical examination, it is important to evaluate the location of the pain over the joint line (medial or lateral), ROM, leg deformity, state of the anterior cruciate ligament (ACL) and patellofemoral (PF) discomfort. Pain should be isolated to one compartment, either medial or lateral, to be indicated for UKA. Assessing knee stability, the Lachman or anterior drawer test can be used to evaluate the integrity of the ACL clinically. Furthermore, varus and valgus stress tests assess the collateral ligaments and amount of correctability of a leg deformity if present.

Traditionally, knee OA is diagnosed on anteroposterior (AP) and lateral weight-bearing radiographs of the knee. Rosenberg et al's views and additional lower leg alignment radiographs are performed as part of the standard radiological work-up of patients with unicompartmental knee OA. This additional 45° posteroanterior flexion weight-bearing radiograph has a high sensitivity and specificity of detecting isolated lateral OA. For evaluation of the patella and trochlear surfaces of the femur, an adequate Merchant view may be helpful in determining gross malalignment and presence of PFOA.

The severity of knee OA is classified according to the Kellgren-Lawrence (KL) Grading System or Ahlbäck classification (table 1). The most limiting aspect of classification based on radiographic imaging is that it detects joint degeneration only in a more advanced stage.

To overcome this limitation, the clinical utility of MRI becomes more important to assess the early detection of OA in the contralateral compartment. Subtle degenerative changes in the subchondral bone, cartilage, abnormalities in the bone marrow, ligaments, menisci, synovium and joint fluid are all well detected with MRI technology.

The radiographic indications for UKA is unicompartmental knee OA (figure 1), with preservation of the contralateral compartment as shown on weight-bearing and valgus/varus stress radiographs. Preoperatively, stress view radiographs could provide information by means of determining correctability of the deformity, ensuring maintenance of the contralateral joint space, and indirectly assessing the integrity of the ACL and medial collateral ligaments., , Advocates of stress radiographs require the deformity to be correctable to neutral, with preservation of the contralateral joint space. However, a preoperative MRI is used more often to document the absence of significant degenerative changes in the contralateral or PF compartment.

Several authors, including the Oxford Group, reassessed age above 60 years as a contraindication for UKA surgery. They demonstrated similar survival rates (97.3%) and functional outcomes at 10-year follow-up compared with patients older than 60 years (95.1%). The fear of early polyethylene wear in younger patients, mostly more active patients at that age, is therefore not supported. Interestingly, a trend of better functional outcomes is seen in this group. This may be explained by the fact that younger patients have high activity levels and high functional demands which are met by UKA, including quicker recovery after surgery and wider ROM.

A general increase in the number of obese patients has been noted in orthopaedic practice over the past few decades, and this trend is likely to continue. Reticence in performing surgery on these patients is due to a possibly increased risk of perioperative complications and poor survival due to early implant failure secondary to component loosening and/or excessive wear. This concern may be particularly relevant with UKA, on account of the potential of point loading at the small area of the bone–implant interface. However, Murray et al performed a large retrospective study and divided 2438 patients into the specific subgroups (body mass index (BMI) <25, 25–30, 30–35, 35–40, 40–45, >45 kg/m²). They demonstrated that the survival rate of the Oxford UKA does not decrease with increasing BMI, and no statistical differences were found between any of the groups at the 5-year or 10-year follow-up. Similar results have been found by other authors and systematic reviews as well.

The most contentious potential contraindication relates to the state of the PF joint. According to the Kozinn and Scott selection criteria, PFOA was one of the contraindications for UKA. However, in 1986, Goodfellow and O'Connor performed a bicompartmental study of the Oxford knee in a series of 125 patients and found no relationship between the state of the PF joint, as seen during surgery, and the outcomes. Therefore, the Oxford Group made the recommendation of ignoring the grade of PFOA when deciding whether or not to implant a UKA.

Current literature confirms this by not showing a relationship between preoperative PFOA and inferior outcomes. Beard et al examined 824 consecutive knees, in which 16% had full-thickness cartilage loss at any location in the PF joint. These patients did not report worse outcomes than those with a normal or near-normal joint surface. Recent reports suggest that this might be the result of indirect PF joint congruence improvement as a result of medial UKA implantation. By restoring the alignment, the contact forces over the PF joint are lowered. Despite the lack of level I evidence, these previously mentioned studies all suggest that PFOA does not influence UKA outcomes. , , ,

From a historical perspective, it was generally accepted that UKA is contraindicated if the ACL is functionally deficient. The first reports highlighted a higher incidence of complications following UKA surgery in ACL-deficient knees, in terms of tibial loosening and a higher revision rate. Mancuso et al summarised the evidence in the literature concerning ACL deficiency in UKA surgery; they concluded that combining ACL reconstruction and UKA is the preferred treatment option for patients with ACL deficiency and bone-on-bone medial OA. Simultaneous or staged ACL reconstruction tends to provide superior outcomes, in particular in younger and more active patients. In the elderly, UKA without ACL reconstruction seems to be a reasonable and attractive option if a fixed-bearing design is used, but careful patient selection is necessary. The literature shows no statistical difference between survival rates of UKAs implanted in ACL-deficient and ACL-intact knees. However, a cautious approach is required, since long-term results are lacking.

Chondrocalcinosis, deposition of calcium pyrophosphate crystals in fibrocartilage and hyaline cartilage, is commonly seen in knees with OA. It is believed that chondrocalcinosis leads to a more aggressive form of OA, potentially leading to accelerated contralateral compartment OA following UKA. Despite the limited number of series, the literature does not support this theoretical disadvantage. Hernigou et al proved the incorrectness of this theory; only 11% of their patients showed progression of OA of the other compartment, which is equivalent or less than UKA knees without chondrocalcinosis. Another report by the Oxford Group showed no significant difference in survival between patients with radiological chondrocalcinosis undergoing medial UKA and controls without chondrocalcinosis. The relevance of histological chondrocalcinosis in patients with UKA remains unclear. Although it is associated with a significantly higher revision rate, these patients report significantly better functional outcomes.

To summarise, over the past two decades the original contraindications to performing UKA surgery have been reassessed by multiple investigators and now the current literature would suggest that age, BMI, PFOA, chondrocalcinosis and ACL integrity are not absolute contraindications for UKA.

Initially, both cemented and cementless designs were used. However, the cementless designs were less reliable with failure rates up to 20% 10 years after surgery. Cementation has proven to be an adequate fixation method for UKA and is therefore considered the standard technique. It has shown high survivorship rates and good functional outcomes. The most common cause of failure of the cemented implant is aseptic loosening according to the joint registries and large systematic reviews. Errors in cementation, thermal necrosis, misinterpretation of radiolucent lines (RLLs), and formation of fibrocartilage and fibrous tissue at the bone—cement interface could all contribute to loosening of the cemented UKA. As a result, a resurgence of interest in cementless fixation has been noted over the past decade to address these perceived disadvantages of cemented fixation.

Modern advances, such as the use of porous titanium and especially hydroxyapatite coating, are responsible for an improved fixation of the cementless UKA. Osseous stability, either by ingrowth or ongrowth, and press-fit fixation of both components are key elements in cementless fixation. Currently, the Oxford UKA is the most commonly used cementless prosthesis. The possible downside of the press-fit fixation is an increased risk of periprosthetic fractures, particularly on the tibial side in older osteopaenic women. More impaction is required to introduce components with good primary fixation in cementless replacement. Despite early conflicting results, recent evidence shows good results on the effectiveness and safety of cementless UKA in mid-term follow-up with randomised controlled trials and case series. Summarised in a recent systematic review by Campi et al, the cementless technique has many advantages in comparison to cemented UKA, including shorter surgical time, avoidance of cementation errors, lower incidence of RLLs and reliable fixation. Despite these promising results, longer follow-up data are required to assess the long-term advantage of cementless UKA.

The first available UKAs were fixed-bearing designs, which often had a flat tibial articular surface. These were less conforming as flexion occurred, and therefore led to higher point loading on the surface. As a result, higher stress within the polyethylene were noted, which increased the risk of component loosening and polyethylene wear. In order to minimise polyethylene wear, Goodfellow and O'Conner designed a mobile-bearing metal-backed UKA in 1986. The articulating surfaces of the components are congruent over the entire ROM in most mobile-bearing designs. Large contact areas and small contact stresses diminish the likelihood of wear and decouple the forces at the implant bone interface, which should reduce the incidence of aseptic loosening. Stability of the insert is created by ligamentous tension and, to a much lesser extent, by the components itself. Therefore, it is mandatory to produce equal flexion and extension balance to maintain stability and reduce the risk of bearing dislocation. Impingement of the mobile-bearing insert is another complication inherent to mobile implants and careful assessment intraoperatively of bearing tracking should alleviate this problem.

Bearing dislocation was observed more often in lateral UKAs (11%) with mobile-bearing designs, caused by a more lax lateral compartment in flexion compared with a tighter medial compartment. This allows the lateral compartment to be distracted by about 7 mm, compared with 2 mm on the medial side. To overcome this problem, the Oxford Group developed a new lateral mobile-bearing tibial component. The Domed Lateral Oxford UKA (Biomet UK) has a spherically convex and domed tibial plateau. Additionally, the biconcave bearing has a 7 mm entrapment anteriorly and posteriorly in order to reduce the likelihood of dislocation. Survival rates of lateral UKA increased up to 92% at a mean follow-up of 4 years and good functional outcomes were reported by Weston-Simons et al. Comparative studies on medial UKAs were performed by Parratte et al and Whittaker et al; they found equivalent mid-term and long-term functional outcomes and survivorship rates of mobile-bearing versus fixed-bearing implants. The predominant reasons for revision were progression of OA and aseptic loosening in both fixed-bearing and mobile-bearing UKA. Similar findings were reported by the national arthroplasty registries, suggesting no conclusive advantage of one bearing design over another., ,

Historically, two fixed-bearing designs have been used for tibial resurfacing when performing a UKA: (1) inlay (all-polyethylene) and (2) onlay (metal-backed). Inlay components are all-polyethylene implants cemented into a carved pocket on the tibial surface, thereby relying on the subchondral bone to support the implant. Onlay components commonly have a metal base plate and are placed on top of a flat tibial cut, supported by a rim of cortical bone. Walker et al used a biomechanical model to compare inlay versus onlay implants, and showed superior load distribution over the tibial surface for the metal-backed onlay design. It has been suggested that this may be a mechanistic explanation for the improved pain relief demonstrated by the onlay components. An additional benefit of metal-backed tibial trays is the possibility to apply cementless fixation. However, metal-backed designs allow a less conservative tibial cut when compared with all-polyethylene implants. In order to minimise contact stresses in the tibial component, a polyethylene thickness of 8 mm should be pursued when possible. Taking into account the thickness of the polyethylene and the metal tray itself (3–4 mm), metal-backed designs necessitate a larger tibial cut. In current practice, metal-backed as well as all-polyethylene tibial implants are being used. The metal-backed design may favour of the renewed interest of cementless fixation.

Conventional manual techniques have been routinely used in UKA surgery with implant position and alignment critical to short-term and long-term outcomes. These variables are most often manually controlled with the aid of extramedullary and intramedullary alignment guides. Although national registries reported lower rates, a recent systematic review showed a 10-year survivorship of medial and lateral UKA of 92% and 91%, respectively., , , , As is described, the accuracy of implant alignment is an important prognostic factor for long-term implant survival; therefore, tight control is recommended.

Over the past decade, there has been a growing interest in surgical quantifiable variables that can be controlled intraoperatively, which include lower leg alignment, soft-tissue balancing, joint line maintenance and component alignment., , , Technical innovations in UKA surgery have led to the development and usage of computer navigation systems, with the purpose of more accurate and tight control of the aforementioned surgical factors., , Meta-analyses have reported improvement of alignment and surgical cutting accuracy, however, failed to show the superiority of functional outcomes in comparison to conventional techniques. As a result, robot-assisted systems have been developed to control these variables intraoperatively and, in addition, refine and enhance the accuracy of the procedure. The fundamental goals of robotic-assisted surgery are to be patient-specific, minimally invasive and highly precise. Most importantly, the robotic systems are ‘semiactive’, meaning that the surgeon retains ultimate control of the procedure while benefiting from robotic guidance within target zones and surgical field boundaries. Preoperative CT-based planning was essential in earlier systems; however, new technology allows image-free robotic assistance (figure 2). Through mapping condylar landmarks and determination of alignment indices, the volume and orientation of bone to be removed is defined. Continuous intraoperative visual feedback provides quantification of soft-tissue balancing and component alignment (figure 3). Compared with conventional UKA, robotic-assisted systems have demonstrated improved surgical accuracy, lower leg and component alignment. , , Another benefit in the use of the robotic system may be a shorter or rapid progression up the learning curve, which can minimise failures related to surgeon workload.

Cobb et al performed a randomised control trial to compare conventional techniques with robot-assisted surgery on 27 patients with medial UKA. They found that the robotic-assisted group had a mechanical axis within two degrees of neutral, while only 40% of the conventional group was in that range. Furthermore, they assessed functional outcomes according to the Western Ontario and McMaster Universities Arthritis Index (WOMAC) score and noted a trend towards improvement in performance with increasing accuracy at 6 weeks and 3 months postoperatively. The optimal alignment for medial UKA is between 1° and 4° varus; this was associated with a better outcome and medium-term to long-term survivorship. For lateral UKA, valgus alignment of 3–7° was correlated with the best functional outcomes at 2 years postoperatively. Pearle et al reported the preliminary results of a multicentre study of 854 patients and found a survivorship of 98.9% and satisfaction rate of 92% at a minimum 2-year follow-up. Comparing these results to other large conventional UKA cohorts may suggest that robotic-assisted surgery could possibly improve survivorship at short-term follow-up.

Drawbacks of robot-assisted surgery are high overall costs and radiation; however, the implementation of image-free robotic assistance has significantly decreased the radiation by eliminating the CT preoperatively. Furthermore, Moschetti et al has shown that robot-assisted UKA is cost-effective compared with conventional UKA when the annual case volume exceeds 94 UKAs per year. Another disadvantage in comparison to conventional techniques is the necessity of pin tracts for the required optical tracking arrays, which is necessary for some robot-assisted systems. They could create a stress riser in the cortical bone when the pins are applied. Nevertheless, prospective clinical studies with longer follow-up are required to assess the additional value of robotic-assisted UKA surgery, despite the promising short-term results.

A French multicentre study of 418 failed knees concluded that aseptic loosening was the most common cause of failure in their population, accounting for 44% of all cases. Similar findings were shown by van der List et al and Citak et al; both reported aseptic loosening and progression of OA as the most common modes of failure in medial UKA. Tibial loosening was seen more often than femoral loosening; moreover, it developed significantly earlier (37.7% within 2 years) when compared with femoral loosening. Noteworthy is the fact that aseptic loosening is much more common in medial than lateral UKA.

Progression of OA in the contralateral compartment accounts for the second most common cause of failure of UKA. Various studies have reported progression of the underlying disease in up to 36% of the knees. To minimise this progression, a high level of accuracy is required for optimal positioning of the components and restoration of the joint line. The restoration of the prosthetic joint space affects load transfers between the two femorotibial compartments. To that end, Khamaisy et al proved a significant improvement of the congruity of the contralateral compartment following medial and lateral UKA implantation. Restoration of the appropriate joint line in the damaged compartment has an influence on survivorship. A joint space height difference <2 mm was significantly associated with shorter medial UKA survival. Failures related to a lower position of the prosthetic joint line were due to loosening, whereas failures related to a higher position of the prosthetic joint space were due to early polyethylene wear and progression of OA in the contralateral compartment. As Chatellard et al stated, UKA acts as a wedge that compensates for the joint damage, which restores normal kinematics and blocks the vicious circle of medial femorotibial OA.

As previously mentioned, wear is another mode of failure which is mostly seen in fixed-bearing designs of UKA. Higher stresses are generated in these types of designs, often in combination with a metal-backed tibial tray, which allows only a certain polyethylene thickness. Thinner polyethylene is at risk for accelerated wear of the increased contact stresses. Furthermore, leg alignment and the position of the components influences wear in the knee following medial UKA. Hernigou and Deschamps showed that a varus undercorrection was associated with increased polyethylene wear and recurrence of the deformity. Subsequently, the risk of lateral degeneration was increased in case of valgus overcorrection. In contrast, no significant correlation was found between polyethylene wear and BMI, gender or preoperative diagnosis of the patient.

Unexplained pain is an important source of failure following UKA surgery. Among 4–23% of the patients with UKA experience pain postoperatively without any obvious reason after the traditional examinations. Park et al recently performed a diagnostic MRI-based study, in order to create a greater insight into the aetiology of the symptomatic patients where physical and traditional radiographs were not aberrant. MRI examination was found to be instrumental in diagnosing these patients. The most common pathologies based on MRIs included loose bodies, osteolysis, tibial loosening, synovitis, stress fractures and infection. Baker et al compared the proportion of UKA and TKA revisions that were performed because of unexplained pain as recorded in the National Joint Registry of England and Wales. The risk of revision was greater following UKA, and proportionally more unicompartmental implants were revised for unexplained pain. Some potential explanations were suggested by the authors. First, UKA revision is perceived as an easier procedure to revise than a TKA and this is likely to lower the threshold of patient and surgeon to proceed with pain as the only indicator. However, registry-based studies have shown that revising a UKA results in a poorer result than a primary TKA, with survival and patient-reported outcomes similar to revising a TKA. They conclude that a demonstrable cause for the revision, rather than unexplained pain, should be the reason for conversion to TKA. Second, inexperienced surgeons faced with an unhappy patient with a UKA with no obvious diagnosis are more likely to blame the unresurfaced compartment. This situation is similar to TKA with an unresurfaced patellar, where the patellar is subsequently resurfaced as it is assumed that the pain must be coming from this articulation. Revision procedures in these patients only result in 25% satisfaction rates, even in the presence of a ‘hot’ nuclear bone scan.

Aseptic loosening, progression of OA, polyethylene wear, bearing dislocation and unexplained pain are the most common failure modes following UKA surgery. To a much lesser extent, instability, infection, malalignment, fracture and tibial subsidence are reported as a cause of failure in current literature.