Imageless navigation system does not improve component rotational alignment in total knee arthroplasty

Abstract

Background: The aim of computer-assisted surgery is to improve accuracy and limit the range of surgical variability. However, a worldwide debate exists regarding the importance and usefulness of computer-assisted navigation for total knee arthroplasty (TKA). The main purpose of this study is to summarize and compare the radiographic outcomes of TKA performed using imageless computer-assisted navigation compared with conventional techniques.

Materials and methods: An electronic search of PubMed, EMBASE, Web of Science, and Cochrane library databases was made, in addition to manual search of major orthopedic journals. A meta-analysis of 29 quasi-randomized/randomized controlled trials (quasi-RCTs/RCTs) and 11 prospective comparative studies was conducted through a random effects model. Additional a priori sources of clinical heterogeneity were evaluated by subgroup analysis with regard to radiographic methods.

Results: When the outlier cut-off value of lower limb axis was defined as ±2° or ±3° from the neutral, the postoperative full-length radiographs demonstrated that the risk ratio was 0.54 or 0.39, respectively, which were in favor of the navigated group. When the cut-off value used for the alignment in the coronal and sagittal plane was 2° or 3°, imageless navigation significantly reduced the outlier rate of the femoral and tibial components compared with the conventional group. Notably, computed tomography scans demonstrated no statistically significant differences between the two groups regarding the outliers in the rotational alignment of the femoral and tibial components; however, there was strong statistical heterogeneity.

Conclusions: Our results indicated that imageless computer-assisted navigation systems improve lower limb axis and component orientation in the coronal and sagittal planes, but not the rotational alignment in TKA. Further multiple-center clinical trials with long-term follow-up are needed to determine differences in the clinical and functional outcomes of knee arthroplasties performed using computer-assisted techniques.