Direct measurement of malrotation of traumatic femoral neck fractures after osteosynthesis: Introduction of a novel method and interrater reliability.

BACKGROUND

In elderly patients, displaced femoral neck fractures are mostly treated by arthroplasty; however for younger patients (<50 years), open reduction and internal fixation is considered the gold standard approach. Despite there being no consensus on the specific procedure, everyday clinical practice in a level I trauma center has shown that postoperative maltorsion after internal fixation of femoral neck fractures can have a significantly worse impact on mobilization and outcome. Different methods for measurement of malrotations are reported in literature. However, any reported method for the assessment of a shaft malrotation in the femur does not work here. In femoral neck fractures, the pointer function of the femoral neck, which is absolutely essential for these techniques, is lost and cannot be set in relation to the condylar plane. These circumstances are not addressed in literature thus far. Therefore, we propose here a novel method to fill this diagnostic gap.

METHODS AND FINDINGS

Three investigators (1 orthopaedic surgeons and 2 radiologists) measured the torsion of 20 legs on 10 patients using the Jarret method and a new geometric technique. To determine the intraobserver reliability the torsional angles were calculated again after 3 months. We applied a new geometric technique, without the need to include the femoral condyles in the measurement, to directly measure the angulation. For torsional difference, the interrater reliability -ICC (interclass correlation) between all investigators was 0.887 (good) (significance level: 95%CI, 0.668-0.969; p<0.001), by using the method of Jarret et al. and 0.933 (good) for the novel technique (significance level: 95%CI, 0.802-0.982; p<0.001). If the examinations are classified according to the patient side, our data show that for established methods, an ICC between the examiners on the right lower extremity is 0.978 (good) (95%CI, 0.936-0.994; p<0.001) and that on the left extremity is 0.955 (good) (95%CI, 0.867-0.988; p<0.001). Comparing with the new method, the right side assumes an ICC of 0.971 (good) (95%CI, 0.914-0.992; p<0.001), while the left side assumes an ICC of 0.910 (good) (95%CI, 0,736-0.976; p<0.001). When it comes to the intraobserver reliability, the measured cohort shows a significant better ICC for the novel method compared to Jarrett et al, with 0.907 respectively 0.786 for comparison in torsional differences.

CONCLUSION

The established methods may fail in assessing this special aspect of malrotation after femoral neck fractures. Here, the method presented results in a significant difference between the injured and uninjured side and shows significant differences in results compared to conventional measurement methods. The inter- and intraobserver reliability determined in this study is excellent and even higher in the assessment of torsional differences than the established method. We believe that the measurement method presented in this study is a useful tool to objectify the postoperative deformities in this area and making therapy recommendations in the future.