Rethinking the Value of Computer-Assisted Surgery

Abstract

Commentary In this study from the New Zealand Joint Registry (NZJR), the authors compared 2 cohorts of patients, 1 managed with computer-assisted surgery (CAS) and the other managed with conventional manual instruments. The investigators held constant the device being used and the 2 instrumentation options, compared only high-volume surgeons, and adjusted for demographic variables in order to minimize selection bias. The preferred alignment philosophy was not reported, nor was actual alignment. The statistical methodology was appropriate, and the authors clearly acknowledged the limitations of the study. The results demonstrated equivalent (and enviably low) revision rates and patient-reported outcome measures (PROMs) in both groups. One can take minor issue with the choice of the Oxford Knee Score (OKS), which is understood to have a ceiling effect1, and still appreciate that there was no meaningful difference in this particular clinical outcome score between the 2 groups. This study is not the first that has failed to show an improvement in clinical results and revision rates when CAS has been used during total knee arthroplasty (TKA). Yet, there is general agreement that CAS (1) increases alignment accuracy and (2) does not produce worse outcomes. One reason may be that there simply is not much room for improvement when the reported results associated with traditional instrumentation are already excellent. In classical teaching, it is generally proposed that variance from mechanical alignment will lead to an increased risk of implant failure. Indeed, the perceived negative impact of malalignment (defined as any tibial alignment not orthogonal to the mechanical axis) is the principal reason why CAS was invented in the first place. Concerns about the effect of alignment on the clinical outcomes of TKA have their origins in the earliest implementations of the surgical technique2. Early devices with flat-on-flat articulations, polyethylene sterilized in air, problematic condylar designs, unreliable cement compounds, vaguely defined surgical goals, and inaccurate instrumentation were very susceptible to off-axis loading caused by misaligned implants. Indeed, the most vexing variables to control in that era were implant alignment and ligament balancing, both of which were identified as causes of early failure3. Low-volume surgeons and patients with complex anatomy posed the greatest challenge with respect to these 2 variables. Enter into this conversation the concept of CAS, which promised to make every surgeon as good at alignment and balancing as a high-volume surgeon, regardless of the complexity of the anatomy. By and large, the technology worked as advertised, consistently delivering better-aligned knees4,5 than were achieved with traditional instrumentation. Proponents therefore believed that CAS would be associated with lower revision rates than manual instrumentation at intermediate to long-term follow-up. Unfortunately, in the aggregate, studies failed to show consistent improvements in survivorship over traditional instrumentation6. By the time CAS was introduced, the majority of implant designs had already transitioned to more congruent profiles with modern bearing surfaces and effective cementing techniques. These devices were, and remain, far better at handling and distributing loads than their predecessors. As an example, while screening for a research protocol, our laboratory tested strain in the proximal part of the tibia when loading 2 different trays, 1 placed in neutral and 1 placed in 5° of varus. The strain patterns changed markedly between the tibiae that received an older titanium tray and those that received a modern cobalt-chromium tray with a broader keel. The stiffer design effectively neutralized any variation in cortical strain associated with off-axis loading. If new implants can handle greater loads, this factor may help to explain why the accuracy of alignment may no longer be as important to impla