OTA 2002 Posters

Interobserver Reliability of the Young-Burgess and Tile Classification Systems of Fractures of the Pelvic Ring

Henry Koo, MD; Mike Leveridge, BSc; Mohit Bhandari, MD, FRCS(C); Hans J. Kreder, MD, FRCS(C); David J. Stephen, MD, FRCS(C); Michael D. McKee, MD, FRCS(C); Emil H. Schemitsch, MD, FRCS(C); St. Michael's Hospital, University of Toronto, Toronto, Ontario, Canada

Purpose: Accurate and rapid evaluation of pelvic fractures is necessary for this potentially life threatening injury. The Young-Burgess and Tile Classification systems of pelvic fractures are widely used today. The purpose of this study was to measure the interobserver reliability of these classification systems and to determine whether the addition of computed tomography (CT) improves their reliability. The reliability of several pertinent radiographic findings was also tested.

Methods: The records of 30 patients taken from a database at a level I trauma facility were reviewed. For each patient, three radiographs (AP pelvis, inlet, and outlet) and computed tomograms were available. The reviewers consisted of six different orthopaedic surgeons with different levels of expertise (two pelvis/acetabular specialists, two orthopaedic traumatologists, and two senior orthopaedic trainees). Each reviewer classified the injury according to the Young-Burgess and Tile classification systems after reviewing the plain radiographs. Each reviewer then repeated the classification after viewing the CT scans. The kappa coefficient was used to determine interobserver reliability for these two classification systems before and after CT scan. The reviewers were also asked to answer these six questions: is the fracture stable? is posterior displacement >1 cm?, is there avulsion of the sacrospinous ligament? is there fracture of the L5 transverse process? is the symphyseal diastasis >2.5 cm? is there associated acetabular fracture? Once again, the before- and after-CT kappa values were determined for these radiographic findings.

Results:

Plain radiographs: The overall kappa value for the Young-Burgess and Tile classification systems was 0.72 and 0.30, respectively. Among the pelvis/acetabular surgeons, the kappa values were 0.85 (Young/Burgess) and 0.84 (Tile). The orthopaedic traumatologists demonstrated kappa values of 0.68 (Young-Burgess) and 0.07 (Tile). In contrast, the kappa values among the orthopaedic trainees were 0.52 (Young-Burgess) and 0.11 (Tile).

CT and plain radiographs: The overall kappa value for the Young-Burgess and Tile classification systems was 0.63 and 0.33, respectively. Among the pelvis/acetabular surgeons, the kappa values were 0.76 (Young/Burgess) and 0.75 (Tile). The orthopaedic traumatologists demonstrated kappa values of 0.65 (Young-Burgess) and 0.30 (Tile), and the orthopaedic trainees had kappa values of 0.59 (Young-Burgess) and 0.17 (Tile).

Questions: The addition of CT significantly improved the reviewers' interpretation of whether the fracture was stable or not. The before-CT kappa value overall was 0.59, whereas the after-CT kappa value was near perfect at 0.93.

Discussion: The Young-Burgess system demonstrated a higher degree of agreement overall. The kappa values were higher for the Young/Burgess system for each subgroup of observers, particularly among the trainees. Reliability did correlate to the amount of experience of the reviewer. The pelvis/acetabular surgeons demonstrated the highest kappa values for both classifications, and the lowest kappa values were found among the trainees. This trend was observed both before and after the addition of a CT scan. The addition of CT did not influence the overall reliability of either fracture classification system. The only subgroup in which CT improved reliability was among the orthopaedic traumatologists. However, CT did significantly improve the reliability as to whether the injury was interpreted as stable or not.

Conclusion: The Young-Burgess system is a more practical classification system for those surgeons who do not have specialty training in pelvic and acetabular surgery. Computed tomography can help orthopaedic surgeons to determine the best treatment for their patients by reliably determining the stability of the fracture.