The impact of patient-specific instrumentation on unicompartmental knee arthroplasty: a prospective randomised controlled study

The Oxford unicompartmental knee arthroplasty (UKA) is the most commonly performed UKA in England, accounting for 62% of all those performed [8]. Numerous studies have demonstrated excellent long-term survival rates in large cohorts [23, 28, 29]. However, the survival rate of the UKA in joint registries is lower than that achieved in large cohort studies [8]. Registry data have highlighted the role of experience and surgical caseload in determining successful outcomes after UKA, demonstrating better results for UKAs performed by high-volume surgeons in high-volume centres [2, 21]. Technical difficulty of UKA is thought to contribute significantly to the variations observed in surgical performance and results [9, 10, 14, 22, 23, 30, 32, 35]. The implication is that those aiming to undertake UKA must be adequately trained and perform a certain number of cases annually. In addition, particularly for inexperienced and low-volume surgeons, there is a need for technological innovations to improve surgical accuracy. A technological tool that has recently received significant attention is patient-specific instrumentation (PSI) for knee arthroplasty [1, 20, 33]. This uses 3-D imaging techniques (MRI or CT) and rapid prototyping technology to produce patient-specific guides for making the femoral and tibial bone resections. PSI systems aim to improve accuracy of implant positioning in addition to the reduction of fat embolism risk, instrument inventory, and operative time [3, 7, 25, 31, 37]. The PSI system developed for the Oxford UKA is the “Signature” system (Zimmer Biomet Inc, Warsaw, IN, USA). It aims to simplify and improve the accuracy of surgical implantation, which is important in determining the outcome of the procedure [5, 13]. If this PSI system works reliably, it is likely to be particularly useful for low-volume UKA surgeons. However, before this can be done, experienced surgeons should assess the system’s reliability. Recent case-series and laboratory-based experiments have demonstrated that PSI technology can improve component positioning during UKA surgery [11, 15, 18, 36]. However, the effect of this technology has received little attention in randomised studies.

Before inexperienced surgeons use such technology, experienced UKA surgeons must evaluate its safety and reliability in order to ensure that no harm comes to patients. The purpose of this prospective randomised controlled trial (RCT) was therefore to compare the accuracy of implantation and functional outcome of mobile-bearing medial UKAs implanted with and without PSI by experienced UKA surgeons.

Inclusion criteria were standard for medial OUKA:

Exclusion criteria were as follows:

The flow of patients through the trial is presented in Fig. 1. Patients were randomised to either PSI or Conventional Instrumentation (CI) group by one of the investigators (AA). The Oxford Microplasty instrumentation was used to implant all UKAs in the CI group. There were 23 patients (23 knees) in the PSI group and 22 patients (22 knees) in the CI group. Randomisation was performed using sealed opaque envelopes. Blinding of the operating surgeons and patients was not possible owing to the surgeon needing to confirm the PSI plans, and the patients undergoing pre-operative MRI scans. PSI group patients underwent an MRI scan using the protocol outlined by the PSI manufacturers to plan development of the PSI guides. The preliminary plan indicating prosthesis size, positioning, alignment, and proposed bone resection levels was reviewed by the surgeons who accepted the default pre-operative plans unless gross errors were present. The patient-specific cutting guides were then manufactured and sent for sterilisation.

The most important finding of this study was that the OUKA PSI system had similar accuracy to conventional instrumentation with regard to femoral component positioning and alignment, but was less accurate for tibial component positioning resulting in unnecessarily deep tibial resections. Although there was no difference in functional outcome between the two cohorts at one year, we conclude that, in the hands of experienced UKA surgeons, the iteration of the PSI system assessed in the current study was not better than current (conventional) instrumentation.

An important finding, which suggests that the PSI was inferior to the conventional system, was that a significantly lower number of cases in the PSI group had the “optimum” bearing size implanted (3 or 4 mm; 52% of the PSI group compared to 95% of the CI group). The PSI group tended to have thicker bearings, which are associated with higher failure rates in the long term [28]. In addition, we believe that deep resections pose the danger of damaging the deep fibres of the medial collateral ligament and the bone may be weakened with an increase in the risk of pain and fracture. Since this study was undertaken, the PSI algorithm has been altered so that the resections are not so deep. Another important finding was that in some cases, the tibial PSI guides did not register adequately on the native tibia. Thus, further work is needed to improve the tibial PSI guides. The unexpected finding of an increased operative time observed in the PSI group could most likely be explained by the fact that the surgeons were intentionally devoting some extra time to evaluating the PSI guides intra-operatively so that they could provide meaningful feedback to the PSI development engineers.

The radiological assessment of component alignment and positioning was based on the recommendations of the OUKA design manual and the literature [6, 13]. Although no difference was found in the number of outliers between the PSI and CI groups, the use of PSI resulted in improvements in optimum posterior tibial slope and the optimum medial fit of the tibial component. The clinical relevance of these minor improvements is not clear. In contrast to the current study, a laboratory-based study demonstrated that PSI was more accurate than CI and was equivalent to a robot-assisted system [15] and a case-series by Volpi et al. [36]. showed that PSI was “highly accurate in reproducing what the surgeon had planned”. However, more recent studies have not been able to replicate these promising early results. A retrospective series of 30 Oxford “Signature” UKAs demonstrated similar radiological outcomes as the present study, and comparable results to standard instrumentation [17]. The authors also reported difficulties with use of the tibial guides. Furthermore, a recent study utilising CT scans to determine the accuracy of component positioning and alignment using the OUKA PSI system found that there was no agreement between the pre-operative plans and the post-operative component alignment for the femoral component angle in the sagittal and axial plane and for the tibial component angle in the coronal plane [34]. Similarly, recent RCTs comparing PSI guides with conventional instrumentation for implanting TKAs have demonstrated either no additional benefit [19, 38] or detrimental effects on implant positioning when using PSI [27].

With regard to functional outcome, we found no significant difference in OKS improvement. There is little previous work regarding the effect of PSI on functional outcome, but a case-series by Bell et al. [4] which assessed a different UKA PSI system mirrored our results and demonstrated excellent functional outcome (OKS and Forgotten Joint score) at one year. Only one randomised study has evaluated functional outcome following in UKA, using a different PSI system by using 3-D gait analysis, the SF-12, and the Knee Society Score [26]. This study demonstrated no significant differences at one year between the two groups.

This study had some limitations. Firstly, the surgeons performing the cases were highly experienced OUKA surgeons with extensive previous experience of the CI. These experts are therefore likely to have a very low number (if any) of surgical outliers when using the CI. This potential bias is further amplified by the surgical learning curve associated with the use of the PSI. A concerted effort was made by the surgeons to address this issue by familiarising themselves with the PSI in a previous pilot study. Nevertheless, the previous experience of the surgeons with the conventional instrumentation is very likely to have favoured the CI group and hence it would have been difficult to demonstrate more superior surgical accuracy using PSI. Secondly, the radiological assessment was based on coronal and sagittal alignment. Further evaluation of component rotation using CT scans would have been desirable but was limited by resources. Nevertheless, the radiographic parameters were those recommended by the designers of the OUKA and previous studies have validated their use [6, 13]. The current study was powered to detect important differences in component alignment and positioning. It is likely to have been underpowered for detecting clinically important changes in OKS. However, the rationale for the current study can be justified based on patient safety factors. Finally, the current study did not evaluate the long-term implant survival and risk of revision surgery—a factor that which is key in determining the efficacy of PSI technology. PSI technology is an exciting development that has received significant attention over the past decade. However, further evaluation and improvement of the PSI guides used in the current study are necessary before they can be utilised on a regular basis in day-to-day clinical work.

The results of this study demonstrate that high-volume OUKA surgeons can achieve similar results, in terms of component alignment and functional outcome when using PSI guides for OUKA. However, intra-operative assessment of the tibial guides by expert OUKA surgeons suggests that although the early “Signature” PSI system design used for this study can safely be used by experienced surgeons, it is not appropriate for inexperienced surgeons.