FEMORAL NECK FRACTURES and CLASSIFICATIONS RELIABILITY: ARE WE USING REPRODUCIBLE SYSTEMS?

Introduction: Classification systems for femoral neck fractures are very helpful for surgeons in order to choose the correct treatment. The purpose of this study was to verify the reliability and reproducibility in the most used classifications for proximal medial femoral fracture: Garden and Pauwels classifications. Materials and methods: Five surgeons independently classified 82 intra-capsular femur fractures treated with 3 cannulated screws using Pauwels and Garden classifications measured on pre- and post-operatively X-Rays. Inter-observer agreement was calculated for both classifications using the multi-rater Fleiss' kappa; intra-observer agreement from pre-operative to post-operative XRs for Pauwels classification was also calculated with the same method. Results: Only six cases of AVN were registered. The inter-observers agreement based on pre-operative X-rays was moderate for Pauwels classification (κ 0.528) and slightly lower for Garden classification (κ 0.448); but in the simplified Garden classification (type I and II not displaced vs type III and IV displaced) the agreement measured was almost perfect (κ 0.908). The intra-observer agreement for Pauwels classification was moderate (κ 0.456). All the results were statistically significant. Conclusions: This is the first study evaluating the Pauwels angle on both preoperative and postoperative X-rays to assess their reproducibility. Garden classification has a really high reliability and reproducibility in evaluating displaced and not displaced fractures and consequently can be helpful in choosing the correct treatment. Pauwels classification has a lower inter and intra-observer agreement.