Relationship of mechanical factors to the strength of proximal femur fractures fixed with cancellous screws

The decision of whether to attempt screw fixation of a femoral neck fracture is based partly on the estimated strength of the fixed bone/implant construct in relation to the loads it will be required to bear. The goal of this study was to determine in vitro the relation of the following biomechanical factors to the strength of internally fixed femoral neck fractures subjected to cyclic and failure loading: (a) square of the density of cancellous bone in the femoral head, (b) percent comminution of the inferior fracture surface, (c) moment arm of the joint force, or distance from the axis of the joint force to the fracture surface, and (d) orientation angle of the fracture surface in the medial/lateral plane relative to the axis of the femoral shaft. Femoral neck fractures were created in each of 38 fresh cadaveric proximal femora using a dropweight or with a materials testing machine. After sustaining a displaced fracture, fixation was achieved using three cannulated cancellous bone screws. The fixed femur was then subjected to 10,000 cycles of a sinusoidially varying load acting on the femoral head, parallel to the femoral shaft, with an initial peak magnitude of 2.2 times body weight, while the hip was flexed, extended, and rotated to mimic some motions of gait. Muscle loading was not simulated. The magnitude of the peak load decreased as the femoral head displaced during cycling. The mean of the peak load for each cycle over the duration of the test was defined as the average load. Following cycling, the bone/screw construct was loaded to failure in the same direction, and this measurement was termed the maximum load. Average and maximum load were then correlated to the four biomechanical factors using a multiple regression analysis. These factors correlated to a high degree with average force (R(2) = 0.771; p < 0.0001) and to a lesser but still significant degree with maximum force (R(2) = 0.458; p = 0.012), demonstrating that they can be used to estimate the strength of fixation under these loading conditions, The strongest correlation for average force was with fracture angle (p = 0.005) and for failure force was with moment arm length (p = 0.072).