Interobserver and intraobserver reliability assessment of calcaneal fracture classification systems

The aim of the present study was to assess the reliability of commonly used intra-articular calcaneal fracture classification systems and to compare them with the newer AO Integral Classification of Injuries (ICI) system. Forty computed tomography and radiographic images of 40 intra-articular calcaneal fractures were reviewed independently by 3 reviewers on 2 separate occasions and classified according to the Essex-Lopresti, Atkins, Zwipp and Tscherne, Sanders, and AO-ICI classification systems. The reviewers were unaware of the patients' identity and all aspects of clinical care. The data were analyzed using kappa (κ) statistics to assess the intra- and interobserver reliability. The κ values were calculated for Essex-Lopresti (κ = 0.85 intraobserver, κ = 0.78 interobserver), Atkins (κ = 0.42 intraobserver, κ = 0.73 interobserver), Zwipp and Tscherne (κ = 0.40 intraobserver, κ = 0.47 interobserver), Sanders (κ = 0.31 intraobserver, κ = 0.35 interobserver), and AO-ICI (κ = 0.41 intraobserver, κ = 0.33 interobserver). The AO-ICI classification system had levels of reproducibility similar to that of the Sanders classification, currently the most widely used system. The Essex-Lopresti classification demonstrated improved reliability compared with that reported in previous studies. This can be attributed to using sagittal computed tomography images, in addition to the originally described plain radiographs, for assessment. This improvement is relevant because of its accepted prognostic predictability.