PURPOSE:The purpose of this study was to assess the effect on femoral torsion by rotational osteotomies at three different levels as measured in 3D using both the mechanical and the anatomic axes. METHODS:Ten cadaveric lower extremities underwent femoral osteotomies perpendicular to the anatomic axis (AA) at three levels: subtrochanteric, mid-diaphyseal and supracondylar. Parallel pins were placed, one in each femur segment. Computed tomography (CT) was acquired in post-osteotomies neutral position, then post-external rotation of the femur at each osteotomy level. Femurs were returned to neutral rotation between imaging exams. Using 3D CT reconstructions, custom software calculated femoral torsion (angle between the femoral neck axis and the posterior condylar axis in the transverse plane) and pin angle between segments, reoriented to both the mechanical axis (MA) and the AA. Pin angle and torsion change were compared for the three osteotomy locations (regression analysis and ANOVA performed). RESULTS:Two specimens were omitted (inadequate imaging); the remaining eight donors were 55-90 years old (mean: 64 ± 15 years), CT confirmed no bony defects. All three levels of osteotomy demonstrated significant correlations between the amount of rotation at the osteotomy (pin angle change) and the resulting change in femoral torsion (R square range 0.658-0.847). No significant differences were found between osteotomy level in torsion (MA:p = 0.285, AA:p = 0.156) or in pin angle (MA:p = 0.756, AA:p = 0.753). CONCLUSIONS:Performing a corrective rotational osteotomy orthogonal to the AA achieves the desired effect on MA regardless of location. This suggests that a surgeon's osteotomy level choice may be based on other risks/benefits of the various techniques.