LONG LEG RADIOGRAPHS AS A MEASUREMENT TOOL FOR BIOMECHANICAL AXIS IN TOTAL KNEE REPLACEMENT: COMPARISON OF RADIOGRAPHIC AND NAVIGATION MEASUREMENTS

Abstract

It is generally accepted that Hip-Knee-Ankle (long-leg) radiographs are a good measurement tool for biomechanical axis of the knee and they have been used as the outcome measure for many studies. Most of the surgeons recommend having pre operative and post operative long leg radiographs for total knee replacement surgery, although practice is not as common. We studied the biomechanical axis on long-leg films and compared it with computer navigation. The aims were

1. To find out repeatability of measurements of biomechanical axis with inter observer readings on long leg radiographs

2. To compare X-ray biomechanical axis measurements with Navigation values obtained during total knee replacement surgery.

Our institution routinely uses long-leg radiographs for total knee replacement (TKR) surgery both pre- and postoperatively. A series of 209 patients who had navigated TKR between Jan 2007 and 2008 were selected. Stryker and Orthopilot systems for navigation were used. The intra-operative biomechanical axis measurements from the computer navigation files both pre-and post- implant were recorded. The long leg films were measured with a defined protocol from the digital images on PACS system. Centre of the head of femur was taken as the upper point.

For the knee centre the midpoint of a line joining the distal femoral notch centre and upper tibia was used. For the ankle centre midpoint of the upper talar margin was used. An angle between the three points represented the radiological biomechanical axis. To investigate inter-observer error, two observers measured the pre- and postoperative biomechanical axis on long leg radiographs independently on 57 patients. For the inter-observer measurements on 57 patients, the intraclass correlation coefficient was 0.99 for pre-operative radiographs and 0.98 for post-operative radiographs. Maximum difference between the two observers was 2° in four cases. All other cases showed the same readings or 1° difference. There was a strong correlation, which was statistically significant, between the pre-operative radiographic and navigated measurements with Pearson correlation coefficient of 0.810 (p< 0.001). The maximum difference between the radiographic and navigated measurement was 24 degrees. The relationship between the post-operative measurements was weaker but statistically significant with Pearson correlation coefficient of 0.323 (p< 0.001). The maximum difference between the two methods of measurement was larger 15.5.

It can be concluded from this study that biomechanical axis on a long leg radiograph is a repeatable measure with good inter-observer correlation. Although it is statistically significantly correlated with navigated readings, the absolute values may be different with each method. This raises the question of the reliability of long leg radiographs for the prediction of true biomechanical axis. Most of the larger value differences had a fixed flexion deformities (9 – 45 degrees). This can affect the readings on the long leg radiographs and make the deformity look either smaller or bigger. Also, our knee kinematic study has proven that the deformity does not remain the same in flexion and in the extended knee. This could also account for the difference in the readings. Other possible reasons for differences in the pre operative readings: the weight bearing status and the surgical opening of the joint, before taking the pre operative biomechanical axis measurements. Differences in the post operative readings could be attributed to: weight bearing status, time length between navigation and radiographic measurements (6–12 wks), scarring of the soft tissues in the meantime and flexed posture of knee in the early post operative period.