The development of computer-assisted surgery in total knee arthroplasty continues its search for better accuracy in the spatial positioning of prosthetic components and in achieving the best ideal ligament balance. Many studies have underscored the value of computer-assisted navigation in obtaining precise bone cuts in terms of both orientation and location, which would optimize bone resection and thereby fulfill ligament balancing requirements. Yet improving bone cut accuracy can be undermined by positioning errors of the component at the final stage of implantation. The objective of this prospective study was to assess this possible loss of accuracy and to suggest possible solutions to minimize this risk.

A consecutive series of 50 total knee arthroplasties was studied using an imageless computer navigation system. This study compared the spatial orientation of the prosthesis components determined using software (frontal positioning for the femoral component, frontal and sagittal positioning for the tibial component) with the recorded orientation of the corresponding bone cuts, which allowed us to quantify the loss of accuracy of these predefined positions after cutting. Trial and final implant orientation was taken into account. Moreover, the mechanical axes of the lower limb, the trial and then the final prosthesis in place were compared. Two procedures were abandoned in the study and two patient files were incomplete, which left a series of 46 cases (29 females and 17 males; mean age at surgery, 67 years; mean BMI, 31.27).

Bone cut orientation was consistently found to be satisfactory. Frontal orientation of the final femoral component (0.2° valgus) did not differ statistically significantly from the distal femoral cut (0.3° valgus) and from the orientation of the trial femoral component, as was true of the slope of the tibial component (4.8°) versus the tibial cut (6.3°) and the mechanical axis of the lower limb with the trial prosthesis and the final implant. The frontal plane orientation of the tibial component (0.6° varus) differed statistically significantly from the bone cut (0.1° valgus).

Several studies have demonstrated the value of computer-assisted surgery, notably in the accuracy of the bone cuts, confirming the work reported herein. The loss of accuracy observed between the bone cut and the final implantation can only be explained by soft tissues between the prosthesis and the bone cut, unequal cement thickness, an orientation error in the impaction handle when placing the final implant, or a conflict between the prosthetic keel and cortical bone. Better exposure of the tibial plateaus, discontinuation of cement use, and navigated impaction ancillary tools could reduce these errors.

Level IV. Prospective study.