Distal Humerus Traction Radiographs: Is the Interobserver and Intraobserver Reliability Comparable With Computed Tomography?

Objectives: To compare the interobserver and intraobserver reliability of traction radiographs with 2-dimensional computed tomography (2D CT) in distal humerus fracture classification and characterization. Design: Randomized controlled radiographic review of retrospectively collected data. Setting: Academic Level 1 trauma center. Patients/Participants: Skeletally mature patients with intra-articular distal humerus fractures with both traction radiographs and CT scans were reviewed by 11 orthopaedists from different subspecialties and training levels. Intervention: The intervention involved traction radiographs and 2D CT. Main Outcome Measurements: The main outcome measurements included interobserver and intraobserver reliability of fracture classification by the OTA/AO and Jupiter-Mehne and determination of key fracture characteristics. Results: For the OTA/AO and Jupiter-Mehne classifications, we found a moderate intraobserver agreement with both 2D CT and traction radiographs (kappa = 0.70-0.75). When compared with traction radiographs, 2D CT improved the interobserver reliability of the OTA/AO classification from fair to moderate (kappa = 0.3 to kappa = 0.42) and the identification of a coronal fracture from slight to fair (kappa = 0.2 to kappa = 0.34), which was more pronounced in a subgroup analysis of less-experienced surgeons. When compared with 2D CT, traction radiographs improved the intraobserver reliability of detecting stable affected articular fragments from fair to substantial (kappa = 0.4 to kappa = 0.67). Conclusions: Traction radiographs provide similar diagnostic characteristics as 2D CT in distal humerus fractures. For less-experienced surgeons, 2D CT may improve the identification of coronal fracture lines and articular comminution.