The value of intraoperative computer navigation during total knee arthroplasty has been debated for more than a decade. Despite substantial research, contradictory results and reservations about the cost and efficacy of the technology have contributed to the failure of computer navigation to become the accepted standard for establishing limb alignment during knee replacement surgery.
In their paper, Zhang and colleagues report the results of a well-designed, prospective, randomized study that was conducted to compare the short-term clinical and radiographic results of total knee arthroplasty performed with computer navigation with the results of total knee arthroplasty performed with conventional mechanical instruments. The authors attempted to control for as many variables as possible by studying patients with bilateral knee arthritis who were undergoing simultaneous bilateral knee replacement. For each patient, both the order of the surgery and the knee that was assigned to the computer-navigation group were selected by randomization. During the follow-up period, the clinician, radiologist, and patient were blinded with regard to the type of procedure that had been performed on each knee.
The outcomes recorded included overall limb alignment on full-length radiographs; various parameters of prosthetic positioning on both standard radiographs and computed tomography (CT) scans; and improvements in clinical outcomes, as determined by the Hospital for Special Surgery (HSS) score. The primary finding of the study was that the use of computer navigation resulted in significantly better limb alignment on full-length radiographs, with none of the knees in that group demonstrating >3° of deviation relative to the mechanical axis; in comparison, 28% of the knees treated with the conventional mechanical instruments fell outside of this target range. Other findings included a mean operating time that was thirty minutes (55%) longer with the use of computer navigation, equal improvements in the HSS scores at six months with the two techniques, and no differences in femoral component rotation relative to the transepicondylar axis as determined on CT.
Strengths of the study include the prospective, randomized, and blinded design in a series of patients undergoing a bilateral procedure; use of full-length radiographs for determining limb alignment; use of CT scans for evaluating femoral component rotation, which have not been utilized in the majority of prior reports; and inclusion of clinical outcomes. The major limitations include the very short-term nature of the clinical and radiographic results as well as the lack of information about both the number of surgeons who performed the operations and the surgeons’ relative experience with the mechanical instruments and the navigation system that were used. The authors also failed to examine the cost implications of the 55% increase in operating time noted with the use of the computer navigation and whether there were any differences in complication rates.
Unfortunately, although this study was well designed and executed, the information is not new and it does not resolve the debate about the use of computer navigation in knee arthroplasty1. First, another recent, similarly designed study of patients treated with bilateral knee replacement demonstrated contradictory results2. This likely reflects the fact that individual instrument systems, the inherent skill set of the surgeons using the systems, and the level of experience of the surgeons with both the mechanical instruments and the computer-navigation systems may influence the results at different institutions. Second, even if we discount studies that failed to identify improvements in limb alignment with the use of computer navigation and accept the conclusions of a recent meta-analysis by Mason and coauthors3 that neutral limb alignment and prosthetic positioning are more consistently achieved with computer navigation, the value of these findings has been recently questioned4,5. In a 2010 study by Parratte and associates5 (itself the subject of a web commentary in The Journal of Bone and Joint Surgery4), doubts are raised about the importance of achieving neutral limb alignment as a means for maximizing long-term results after knee arthroplasty. While it is too soon to advise against the goal of achieving limb alignment that approximates the neutral mechanical axis, the work by Parratte and colleagues does require us to re-evaluate whether other variables, such as soft-tissue balancing, rotational positioning of the components, and dynamic gait factors, may influence knee function and long-term prosthetic survival more than static limb alignment does1,5.
In summary, the work by Zhang and colleagues was designed to evaluate whether short-term clinical and radiographic outcomes were different between the two knees of individuals who had undergone bilateral knee replacement with one procedure performed with conventional instruments and the other done with computer navigation. The results demonstrate significantly better restoration of neutral limb alignment when computer navigation was used but no differences in short-term outcomes. These improvements were obtained at the expense of a 55% increase in operating time. Although the authors are to be commended for executing a well-designed study, now, more than at any time in the past decade, the question remains whether the ability of computer navigation to help surgeons achieve neutral limb alignment has any short or long-term clinical advantages.