Elimination of the Femoral Neck in Measuring Femoral Version Allows for Less Variance in Interobserver Reliability

Abstract

Background and Objectives: Producing consistent measures of femoral version amongst observers are necessary to allow for an assessment of version for possible corrective procedures. The purpose of this study was to compare two computed tomography (CT)-based techniques for the reliability of measuring femoral version amongst observers. Materials and Methods: Review was performed for 15 patients post-femoral nailing for comminuted (Winquist III and IV) femoral shaft fractures where CT scanograms were obtained. Two CT-based techniques were utilized to measure femoral version by five observers. Results: The mean femoral version, when utilizing a proximal line drawn down the center of the femoral head-neck through CT, was 9.50 ± 4.82°, while the method utilizing the head and shaft at lesser trochanter centers produced a mean version of 18.73 ± 2.69°. A significant difference was noted between these two (p ≤ 0.001). The method of measuring in the center of the femoral head and neck produced an intraclass correlation coefficient (ICC) of 0.960 with a 95% confidence interval lower bound of 0.909 and upper bound of 0.982. For the method assessing version via the center of the head and shaft at the lesser trochanter region, the ICC was 0.993 with a 95% confidence interval lower bound of 0.987 and an upper bound of 0.996. Conclusions: The method of measuring version proximally through a CT image of the femoral head–neck versus overlaying the femoral head with the femoral shaft at the most prominent aspect of the lesser trochanter produces differing version measurements by roughly 10° while yielding an almost perfect interobserver reliability in the new technique. Both techniques result in significantly high interobserver reliability.