Femoral neck-shaft angle (NSA) was measured in a series of anteroposterior
(AP) hip radiographs of a cadaveric femur in varying degrees of rotation. A
mathematical model was developed to predict NSA on an AP radiograph in var
ying degrees of femoral rotation. The predictions of the model were found t
o correlate well with the experimental data (correlation coefficient = 0.94
). Based on the mathematical model, a wide range of patient positioning was
found to result in a <10<degrees> error in the measurement of femoral NSA.
Coxa vara, increased femoral anteversion, cerebral palsy, and developmenta
l dislocation of the hip resulted in a more restricted range of acceptable
femoral positioning. External rotation of the femur should be avoided durin
g patient positioning because as little as 7 degrees can cause a >10 degree
s change in the apparent NSA. For all patient populations, internally rotat
ing the femur will allow for determination of the femoral NSA to within 10
degrees.