Gunston’s original report of the polycentric knee included follow-up of between 1 and 2.5 years for 22 patients. All 22 patients reported improvements in pain, with an average range of motion of 8.4–101°. All but one patient were recorded having increased levels of mobility. Three knees required manipulation under anaesthesia for ‘delayed healing’. One knee was arthrodesed due to lack of functional improvement (this was on a background of previous knee surgery, with suboptimal prosthesis placement, resulting in an unstable knee) [22]. 10-year follow-up of polycentric knees was reported by Lewallen et al. in 1984, for 209 knees. At 10 years only 42% were assessed as being ‘successful’ – defined as an ability to mobilise without aids, with mild discomfort, and did not require medical attention. A further 24% were reported as successful before 10 years, due to death or loss to follow-up. Thirty-four percent of the knees were classified as failures, with causes reported as instability in 13%, loosening in 7%, infection in 3%, patellofemoral pain in 4% and ankyloses in 2%. At time of revision surgery, loosening was found to be present in an additional 5%. Of note, failure rates were doubled in patients with components implanted in any degree of varus alignment or greater than 8° valgus alignment [30].

Skolnick et al. reported the 2-year outcomes after 119 geometric TKAs (with eight lost to follow up). As with the polycentric knee, pain relief was reported in the majority of patients (92%), an increased walking distance and a reduced requirement for walking aids with a 93% satisfaction rate. Range of motion in patients clinically reviewed was 7–87°. Post-operative flexion was reported as not significantly different to preoperatively, although there was a significantly different increase in extension achieved post-operatively. A deep infection rate of 1.8% was reported. 11.8% demonstrated radiographic tibial component loosening, with 9.1% requiring reoperation because of this. Approximately 80% demonstrated lucent lines at the cement-bone interface [31]. A 2–3.5-year follow-up of several TKA designs was reported by Insall et al. For the Geomedic (50 knees), the average range of motion was 90°, and an average increase in the Hospital for Special Surgery knee rating score of 69% and 85% was reported for osteoarthritis and rheumatoid arthritis patients, respectively. Eleven cases (nine of which were rheumatoid arthritis patients) were considered as failures due to one dislocation, two late infections, two patients with patellofemoral pain, two cases of tibial loosening and two with restricted range of motion. Of note, the authors report a radiolucent line in 80% of the geometric knees, although only 8% as loose, similar to Skolnick et al. [32]. A failure rate of 18.3% at 8.5-year follow-up was reported by Riley & Woodyard, on a cohort of 71 Geomedic TKAs, in which failure was defined as severe pain or need for reoperation [33]. Van Loon et al. reported 70% of knees remained painless at an average of 11-year follow-up and an average range of motion of 100°. A failure rate of 18%, most commonly due to tibial loosening, was similar to that of Riley & Woodyard. A survival rate (with an endpoint of implant removal) of 78% was given at 13 years; however this dropped to 58% if radiographically loose prostheses were included [34].

Ranawat et al. reported their experience of the duocondylar knee with a 2–4-year follow-up in 94 knees. Pain relief of 88.7% was achieved, although only 40.2% were considered excellent or complete pain relief. Average range of motion was 102°. Radiographic lucencies were found in 76% at 3 years, with 26% demonstrating progression. Five to seven patients required revision surgery for loose tibial components [35].

Townley reported his experience of 532 anatomic total knees, with a follow-up period of 2–11 years. An excellent or good outcome was found in 89%, in which the range of motion was beyond 90° and pain or activity restriction was mild to none, and there was no requirement for walking aids. A 2% rate of tibial loosening was reported [36]. A 23-year follow-up of the Townley Anatomic Knee by Pritchett demonstrated an 89% survivorship at 23 years, with revision for any reason as the endpoint. Goniometer measured flexion increased from a pre-operative mean of 104° to 117° post-operatively. Knee Society Scores increased from a pre-operative mean of 42 to a post-operative mean of 91. 5.6% required revision, with polyethylene wear being the most common cause. Tibial loosening was reported as rare [37].

Cloutier et al. has reported the 9–11-year follow-up of 107 Hermes 2C knee. This demonstrated a good or excellent outcome in 97%. The average range of motion was 107 ± 12.6°, normal anteroposterior stability in 89%, average knee score of 91 ± 8.4 and average functional score of 82 ± 21. With an endpoint of revision, the survival rate was 95 ± 2% at 10 years. No radiolucent lines were seen in 91%. Four percent of the knees (from the original cohort of 163) were revised, 3 for deep infection, one for instability after ACL rupture in a rheumatoid arthritis patient, one for a loose femoral component and two revisions for polyethylene wear [20]. The 22-year results for this same group have also been reported, although this constituted only 20% of the original cohort of 163. This demonstrated an average flexion of 103°, knee society score of 87 and function score of 68. With revision for any reason as an endpoint, survival at 22 years was 82.1%. This increased to 96.1% when aseptic loosening was chosen as the endpoint [38].

Such results highlight the difficulties with the earliest designs of BCR TKA with regard to implant loosening, particularly of the tibial components. The 10–20-year follow-up studies of more recent designs have much improved on these, however, and are more comparable to the widespread CR and PS TKAs. Patient satisfaction however, rather than survivorship, is another important factor to consider. A prospective randomised trial, in which patients underwent stage bilateral TKA, with different prostheses, supports BCR TKA. In 440 patients, with a minimum of 2-year follow-up, 89.1% preferred a BCR TKA in one knee to a PS TKA in the other. Also implanted in this study was a medial-pivot knee design, and this was preferred equally to the BCR TKA [39]. The reasons for the preference of a BCR TKA to a CR or PS TKA may be due to the more normal kinematics achieved with a BCR TKA. This has been demonstrated with in vivo fluoroscopic kinematic analysis. The BCR TKA group on deep knee bend demonstrated more normal posterior femoral roll back, compared to a CR TKA, which demonstrated anterior femoral movement on flexion [40]. Anteroposterior laxity has also been shown to be more near normal in BCR TKA than CR TKA in a separate study [41].

Encouraging long-term results from more recently designed BCR TKAs, coupled with patient preference, and evidence of more normal knee kinematics argue that BCR TKA is a viable treatment option. With modern developments, BCR TKA may prove to be the superior option in appropriately selected patients and may help to reduce 10–20% of unsatisfied patients.

However there remain both technical and design challenges to be addressed in BCR TKA. Implantation of the prosthesis is technically demanding and has a notable learning curve in our experience. The inability to sublux the tibia anteriorly with an intact ACL reduces visualisation of the proximal tibia, potentially making resection, templating and implantation more difficult. Concerns remain regarding tibial island fracture or ACL avulsion; however, in experienced hands, rates of tibial island fracture are below 2% [37, 42]. On a similar theme is the issue of knee stiffness, where inadequate bony resection can result in a stiff knee with a reduced range of movement. Minimisation of tibial island fracture and stiffness requires a clear understanding of soft tissue balancing of the knee and the impact of femoral and tibial resection.

The femoral component in BCR TKA is very similar, or identical in some cases, to manufacturers’ existing CR or PS systems. This is in contrast to the tibial component, which to retain the ACL insertion necessitates a cut-out in the plate to produce a U-shape to fit around the tibial island. This results in a narrow connecting beam anteriorly, with resulting concerns regarding fatigue failure. In response to this some manufacturers have used alternative manufacturing techniques or alloys with higher resistance to fatigue failure. More significantly concerns exist regarding tibial tray fixation and loosening. As discussed above several of the early designs suffered from early loosening, and newer designs often demonstrate radiolucent lines in short-term follow-up of unclear significance [32, 42]. The challenges arise again from the inability to sublux the tibia. This reduces the space available to insert the tibial component and so constrains the size of fixation pegs, or keels, that can be used. In comparison to standard TKA designs, BCR TKA tibial trays have significantly smaller fixation features or have required screw fixation. This is coupled with reduced access during application of cement and component impaction. Whilst progressive radiolucent lines are clearly a cause for concern, the implications of ‘stable’ radiolucent lines in the short-term are unclear. It may be that similarly to the Oxford unicompartmental knee, these radiolucent lines are common and do not herald failure [43]. The rates of radiolucent lines also differ between designs, for example, Cloutier et al. reported a rate of 9% at 9–11 years, versus 30% by Christensen et al. at an average of 18-month follow-up [20, 42]. This suggests that tibial tray designs can have a significant influence on the formation of these radiolucent lines. The question remains as to whether the presence of radiolucent lines has an impact on outcome or if such outcomes are similarly design specific. Only long-term follow-up studies can answer these questions for the emerging designs. Finally, with the tibial tray consisting of often two separate polyethylene components, the question remains as to whether different sized bearings can be used to allow for additional soft tissue balancing.

With an increasing recognition of the unsatisfied 20% of patients following total knee replacement, and increasing patient expectations for return to function, it is not surprising that interest in BCR TKA has increased. Whilst some may suggest that BCR TKA has been tried before, with limited if any benefits over CR or PS TKA designs, we feel it would be premature to abandon the concept altogether. It is over 50 years since the first designs for BCR TKA emerged, and our understanding of the knee, implants and manufacturing have come a long way. Manufacturers have collaborated with a number of groups, and as such there are several BCR TKAs available. These include designs such as the Zimmer Biomet Vanguard XP, the BioPro Total Knee Original, the Ceraver Hermes 2C and the Smith & Nephew Journey II XR. These designs incorporate features of those designs that have come before them, with incremental changes to geometry as well as new materials.

Using the Vanguard XP as an example, one can see these design steps. The femoral component is based upon the established Vanguard CR knee system with a trochlear groove design to reduce patellar shear stress. The posterior condyle geometry has been altered with the aim of increased flexion without edge loading. Asymmetric condyles also feature, with a larger lateral condyle to allow greater roll back laterally. The tibial tray consists of the typical U-shaped tray of forged cobalt-chromium-molybdenum, with two pegs and two keels. This appears to be an approach recognising the survivorship reported of four pegged tibial components (although not in an ACL-retaining knee) and their experience of the Oxford knee which uses a keel [44]. Another example of modern technologies being incorporated is the use of vitamin E-infused bearings as a potential means to reduce revision due to polyethylene wear (as seen in the longer-term follow-ups in previous prosthesis). These bearings also incorporate compartment-specific geometries, recognising the difference in kinematics between the medial and lateral sides (Fig. 7.4).

However, new prostheses, whilst attempting to address issues of prior designs, are not guaranteed to succeed. Christensen et al. have reported a higher reoperation with a new BCR TKA design in a retrospective review of 78 BCR TKAs compared to 294 CR TKAs at early follow-up [42]. Radiolucent lines were also seen in 30% as previously discussed.