The aim of this study was to quantify the stability
of fracture-implant complex in fractures after fixation. A total
of 15 patients with an undisplaced fracture of the femoral neck,
treated with either a dynamic hip screw or three cannulated hip
screws, and 16 patients with an AO31-A2 trochanteric fracture treated
with a dynamic hip screw or a Gamma Nail, were included. Radiostereometric
analysis was used at six weeks, four months and 12 months post-operatively
to evaluate shortening and rotation. Migration could be assessed in ten patients with a fracture of
the femoral neck and seven with a trochanteric fracture. By four
months post-operatively, a mean shortening of 5.4 mm (-0.04 to 16.1)
had occurred in the fracture of the femoral neck group and 5.0 mm
(-0.13 to 12.9) in the trochanteric fracture group. A wide range
of rotation occurred in both types of fracture. Right-sided trochanteric
fractures seem more rotationally stable than left-sided fractures. This prospective study shows that migration at the fracture site
occurs continuously during the first four post-operative months,
after which stabilisation occurs. This information may allow the
early recognition of patients at risk of failure of fixation. Cite this article:
Trochanteric femur fractures can be classified using the Jensen modification of the Evans’ classification or the AO/ASIF classification. This study compares the reproducibility of both classifications. Furthermore we evaluated the agreement on fracture stability, choice of osteosynthesis, fracture reduction and the accuracy of implant positioning. We used pre- and postoperative lateral and AP radiographs of 50 trochanteric femur fractures. The fractures were classified using both classifications with a three-month interval by five trauma surgeons and five residents. Inter-, and intra-observer variability was analysed using the multi-rater Fleiss’ kappa and the Cohen’ kappa tests. The AO/ASIF classification showed a kappa coefficient for the intra-observer agreement of 0.40 (SE 0.01). After leaving out classification-subgroups, AO/ASIF classification showed a coefficient of 0.68 (SE 0.02) and the Jensen classification a coefficient of 0.48 (SE 0.02). The kappa values of the intra-observer reliability of the AO/ASIF classification with and without subgroups were: 0.43 (SE 0.08) and 0.71 (SE 0.08) respectively. For the Jensen classification the kappa value was 0.56 (SE 0.09). Preoperative agreement on fracture stability and type of implant to be chosen showed kappa values of 0.39 (SE 0.05) and 0.65 (SE 0.04). Postoperative disagreement on the choice of implant was 15% (kappa 0.17, SE 0.08). Kappa values for postoperative fracture reduction and position of the implant were 0.29 (SE 0.09) and 0.22 (SE 0.05), respectively. Both the Jensen classification and the AO/ASIF classification showed poor reproducibility. However, without subgroups the AO/ASIF classification seemed more reliable. This study suggests that the definition of stability of trochanteric fractures remains controversial, which possibly complicates the choice of osteosynthesis. Refinement of the classifications or renewed definition of trochanteric fractures seems to be required.
