Reliability of classification systems for intertrochanteric fractures of the proximal femur in experienced orthopaedic surgeons

Abstract

The aim of this study was to determine the reliability of currently used classification systems for intertrochanteric fractures of the proximal femur, and to determine the reliability of these systems in experienced orthopaedic surgeons. Forty intertrochanteric fractures of the proximal femur were classified independently by five experienced observers using the AO, Evans, Kyle, and Boyd classification systems on two separate occasions 3 months apart. The interobserver and intraobserver variation was assessed using kappa statistics. The level of agreement for classification into AO groups was almost perfect or substantial, and higher than other classification systems. When the fractures were further classified using the AO classification with subgroups, reliability became worse. The current study suggests that the AO classification system with groups can be used more reliably to measure intertrochanteric fractures of the proximal femur than Evans, Kyle, and Boyd classification systems. However, the reliability of the AO classification with subgroups is not satisfactory.