Background: Fracture-dislocations of the elbow, particularly those that involve a fracture through the proximal ulna, are complex and can be difficult to manage. Moreover, current classification systems often cannot discriminate between Monteggia-variant injury patterns and trans-olecranon fracture-dislocations, particularly when the fracture involves the coronoid. The Mayo classification of proximal trans-ulnar fracture-dislocations categorizes these fractures into 3 types according to what the coronoid is still attached to: trans-olecranon fracture-dislocations (the coronoid is still attached to the ulnar metaphysis); Monteggia-variant fracture-dislocations (the coronoid is still attached to the olecranon); and ulnar basal coronoid fracture-dislocations (the coronoid is not attached to either the olecranon or the ulnar metaphysis). The purpose of this study was to evaluate the intraobserver and interobserver agreement of the Mayo classification system when assessing elbow fracture-dislocations involving the proximal ulna based on radiographs and computed tomography scans.Methods: Three fellowship-trained shoulder and elbow surgeons and 2 fellowship-trained orthopedic trauma surgeons blindly and independently evaluated the radiographs and computed tomography scans of 90 consecutive proximal trans-ulnar fracture-dislocations treated at a level I trauma center. The inclusion criteria included subluxation or dislocation of the elbow and/or radioulnar joint with a complete fracture through the proximal ulna. Each surgeon classified all fractures according to the Mayo classification, which is based on what the coronoid remains attached to (ulnar metaphysis, olecranon, or neither). Intraobserver reliability was determined by scrambling the order of the fractures and having each observer classify all the fractures again after a washout period > 6 weeks. Interobserver reliability was obtained to assess the overall agreement between observers. K Values were calculated for both intraobserver reliability and interobserver reliability.Results: The average intraobserver agreement was 0.87 (almost perfect agreement; range, 0.76-0.91). Interobserver agreement was 0.80 (substantial agreement; range, 0.70-0.90) for the first reading session and 0.89 (almost perfect agreement; range, 0.85-0.93) for the sec-ond reading session. The overall average interobserver agreement was 0.85 (almost perfect agreement; range, 0.79-0.91). Conclusion: Classifying proximal trans-ulnar fracture-dislocations based on what the coronoid remains attached to (olecranon, ulnar metaphysis, or neither) was associated with almost perfect intraobserver and interobserver agreement, regardless of trauma vs. shoulder and elbow fellowship training. Further research is needed to determine whether the use of this classification system leads to the application of principles specific to the management of these injuries and translates into better outcomes.
