胫骨平台骨折三柱分型的可信度评价

Three-Column classification for fracture of tibial plateau and the related reliability assessment

目的研究胫骨平台骨折三柱分型不同观察者间的可信度，并与传统Schatzker分型进行比较，证实三柱分型的可靠性。方法2004年12月至2007年3月，对胫骨平台骨折患者进行三柱分型，根据分型选择手术切口和治疗。采用随机数表法从304例（323膝）胫骨平台骨折中抽取50例，选择4位未参与该项手术的工作人员，在规定时间内完成对每例胫骨平台骨折的三柱分型和Schatzker分型的评估。评估结束后使用Kappa值计算4组相关度，以比较评价三柱分型和Schatzker分型的可信度差别。结果323膝胫骨平台骨折中，0柱骨折4膝（1．2％），单柱骨折181膝（56．0％），双柱骨折108膝（33．4％），三柱骨折30膝（9．3％）。在单柱骨折中，外侧柱114膝（63．0％），内侧柱53膝（29.3％），后侧柱14膝（7．7％）。在双柱骨折中，外侧柱合并内侧柱36膝（33．3％），内侧柱合并后侧柱34膝（31．5％），外侧柱合并后侧柱38膝（35．2％）。三柱分型的可信度Kappa值为0．766（0．706＂-0．890），属于基本可信程度；Schatzker分型的可信度Kappa值为0．567（0．513-0．589），属于中度可信程度。结论基于CT和三维重建的胫骨平台骨折三柱分型直观明了，具有较高可信度，建议作为新的分型系统在临床推广和应用。

＂Objective To investigate the reliability ofthe Three-Column classification for fracture of tibial plateau and compared with the conventional Schatzker classification. Methods From December 2004 to March 2007, 304 patients with tibial plateau fracture were treated in our department. The CT scans and 3D reconstruction were performed in all patients before operation. All the patients were classified by the Three-Column classification and then treated with open reduction and internal fixation via the corresponding approaches based on the Three-Column classification. To validate the Three-Column classification, four doc- tors classified 50 patients which were chosen randomly from 323 consecutive fracture cases. Schatzker clas- sification was also classified at the same time. Inter-observer reliabilities of the Three-Column classification and the Schatzker classification were figured out respectively in the form of Kappa value. Results Accord- ing to the Three-Column classification, there were 4 pure compression fractures （1.2%）, 181 one-column fractures（56.0%）, 108 two-column fractures（33.4%） and 30 three-column fractures （9.3%） in all 323 cases. In one-column fracture cases, the numbers of lateral, medial, and posterior column fractures were 114 （63.0%）, 53 （29.3%）, and 14 （7.7%）, respectively. Among two-column fracture cases, the lateral-medial col- umn fractures （33.3%） were 36. While the number of medial-posterior column fractures and lateral-posterior column fractures were 34 （31.5%） and 38 （35.2%）, respectively. There was a moderate inter-observer relia- bility when using Schatzker classification （Kappa=0.567; range, 0.513-0.589）. There was a substantial inter- observer reliability when using the Three-Column classification （Kappa=0.766; range, 0.706-0.890）. Con- clusion The Three-Column classification based on CT and 3D reconstruction is simple and more reliable for fracture of tibial plateau. It is suggested that the new classification system, the Three-Column classification, can be used and popularized in the orthopaedic surgery.