Introduction: The reproducibility of various classification systems for trochanteric fractures is poor. This problem could be related to a lack of readability when using conventional X-rays.
Hypothesis: Computed tomography scanning (CT scan) will improve the interobserver reproducibility of the AO classification for trochanteric fractures.
Patients and methods: This was a prospective, observational, descriptive study following a group of 53 patients with trochanteric fractures. The fractures were evaluated with conventional X-rays, CT scan slices and 3D reconstruction (3DR). The resulting images were blinded and analysed by two observers using two classification systems: AO and Evans modified by Jensen (EVJE). A sample size of 53 was needed to show an improvement in the interobserver reproducibility when deciding the AO classification type with CT scan images. Kappa coefficients were used to measure interobserver reproducibility and agreement; agreement is the degree of consistency in the analysis by one observer who views the same fracture on two different imaging modalities.
Results: The interobserver reproducibility for the AO classification was 0.28 for X-rays, 0.33 for CT scan and 0.28 for 3DR. For the EVJE classification, these coefficients were 0.50 for X-rays, 0.35 for CT scan and 0.47 for 3DR. The agreement rate between the two imaging modalities was between 0.38 and 0.58 for X-rays/CT scan and between 0.79 and 0.86 for CT scan/3DR.
Discussion: The primary objective of this study was not achieved. CT imaging does not improve the interobserver reproducibility of various classification systems for trochanteric fractures. However, by providing images as slices, the complex nature of fractures in this area was revealed. The challenges related to classifying various fractures are not exclusively related to a "readability" problem, but also an understanding and analysis problem.
Level of evidence: Prospective diagnostic study, level III.
Copyright © 2012 Elsevier Masson SAS. All rights reserved.