Reliability of the commonly used classification systems for interprosthetic fractures.

Abstract

An interprosthetic fracture occurs between a hip and knee arthroplasty. There is currently no universally agreed classification. The aim of this study was to determine the interobserver and intraobserver reliability of the most commonly used interprosthetic fracture classifications. Nineteen interprosthetic fractures were classified by four reviewers for inter- and intraobserver reliability. The most commonly used interprosthetic fracture classifications were the Soenen classification, Platzer classification, and Pires classification. Cohen's kappa coefficient was calculated. A moderate interobserver reliability was found for all the classification systems. The Platzer classification had a kappa value of 0.586, the Pires classification 0.499, and Soenen classification 0.489. The intraobserver error was 0.767 for the Platzer classification (substantial agreement), 0.636 for the Pires classification (substantial agreement), and 0.318 for the Soenen classification (fair agreement). This study has demonstrated moderate interobserver reliability and substantial intraobserver reliability for both the Platzer and Pires classifications. This paper would recommend the use of either classification for interprosthetic fractures.