Despite the well-documented improvement in coronal alignment achieved by computer assisted navigation, varying results have been reported for sagittal alignment. Current navigation systems rely on a sagittal femoral mechanical axis identified by the navigation system, but little information is available on the relationship between the sagittal mechanical axis and anatomical axes for intra-operative or postoperative radiographic assessments. We asked whether deviations exist between sagittal femoral mechanical axis and anatomical axes and attempted to identify predictors of the deviations found.

In 100 consecutive patients (200 knees) undergoing TKA, angles between two anatomical axes (the anterior cortical line and mid-medullary line) and two sagittal mechanical axes identified by current navigation systems were measured as proxies of the deviations between them on true lateral radiographs of the whole femur. Correlation analyses and multivariate regression were carried out to identify predictors of deviations.

Significant deviations existed with wide ranges between the anatomical axes and the sagittal mechanical axes. Degree of femoral bowing and femoral length were found to be predictors of deviations between sagittal femoral mechanical axes and anatomical axes.

This study suggests that surgeons applying navigation technology to TKA need to consider deviations between the sagittal femoral mechanical axes and anatomical axes when they intend to place a femoral component at a target sagittal orientation with respect to an anatomical reference.

Application of computer assisted navigation (CAN) has been documented to improve the accuracy of limb alignment and implant positioning. However, a recent study reported that a great deal of disparities occurred between the radiographic and navigational measurements calling the basic argument for application of CAN to TKA into question. In the authors’ practice using CAN for TKA, we have observed consistent disparities between the preoperative radiographic assessments and intraoperative navigational assessments of limb alignment in the coronal plane. A large disparity between radiographic and navigational assessments of limb alignment would be presenting a challenging question whether or not the surgeon can rely on the information provided by the CAN system. We developed a novel method to measure the coronal limb alignment and have found that the radiographic measurements with the novel method remarkably reduce the disparities between the radiographic and navigational assessments of the coronal limb alignment. This study was conducted to document the existence of the disparities between the radiographic and navigational assessments of the limb alignment and the value of our novel method to perform preoperative radiographic measurements of limb alignment.

In 107 TKAs performed using a CAN system (Ortho-pilot: B. Braun-Aesculap, Tuttlingen, Germany), radiographic assessments of coronal limb alignment were assessed using preoperative and postoperative whole limb radiographs taken with weight bearing with two different methods: a standard method, angle between the femoral mechanical axis (the line connecting hip center and the top pint of the femoral intercondylar notch) and a tibial mechanical axis (the line connecting the mid-point between the medial and lateral tibial eminences and the mid-point of the talus dome) and a novel method, the angle between the weight loading line (the line connecting the hip center and the mid-point of the talus dome) and the tibial mechanical axis. A negative value was given to a varus alignment and a positive value to the valgus alignment. During surgery, the coronal limb alignment was measured by the navigation system two different time-points: after registration and after implantation of prostheses. The disparity between the radiographic and navigational assessments was calculated with subtracting the radiographic assessments by the navigational assessments.

The disparity between the radiographic and navigational assessments was significantly smaller with the novel method than with the standard method. The mean difference between the radiographic and navigational assessments of preoperative limb alignment was −6.5o (range: −19 ~ 1) with the standard method and −0.9o (range: −8o to 4o) with the novel method. The mean difference between the radiographic and navigational assessments of the postoperative limb alignment was −1.96 (range: −11 ~ 3) with the standard method and −1.3 (range: −6 ~3).

This study documents that a wide range of disparities occurs between the radiographic and navigational assessments of limb alignment and the amount of disparity occurs in preoperative assessments. Our findings indicate that our novel method to perform the radiographic assessments of limb alignment can be a useful tool to interpret the information intraoperatively given by the navigation system.