Käfer W, Fraitzl C R, Kinkel S, Puhl W, Kessler S
Orthopädische Klinik mit Querschnittgelähmtenzentrum der Universität Ulm, c/o Rehabilitationskrankenhaus Ulm, Ulm.
Z Orthop Ihre Grenzgeb. 2004 Jan-Feb;142(1):33-9. doi: 10.1055/s-2004-818027.
It was the purpose of this prospective study to analyze validity and reliability of three different radiographic classification systems, which are used to assess bone stock loss prior to revision total hip arthroplasty.
Investigation was performed using the classification systems according to 1) Paprosky et al., 2) Hungerford et al., and 3) the German Society of Orthopaedics and Traumatology (DGOT). Twenty-eight consecutive patients with diagnosis of aseptic failure of total hip arthroplasty were enrolled in the study. Interrater agreement was analyzed by evaluating the classifications preoperatively by three investigators with different level of clinical experience using a blinded observer technique. Validity was analyzed by comparison of preoperative radiographic findings (n = 3) and an intraoperative control by visualization and palpation (n = 1). Spearman coefficient of correlation (r(s)) was used to establish levels of agreement among multiple ordinal variables.
Interrater reliability testing using paired comparison between the three investigators revealed non-homogeneous coefficients of correlation (r(s): Paprosky femur: 0.45 - 0.67, acetabulum: 0.38 - 0.63; Hungerford: 0.46 - 0.66; DGOT femur: 0.38 - 0.59, acetabulum: 0.42 - 0.76). Paired analysis of correlation between preoperative and intraoperative findings again showed non-homogeneous coefficients of correlation (r(s): Paprosky femur: 0.59 - 0.68, acetabulum: 0.39 - 0.70; Hungerford: 0.39 - 0.74; DGOT femur: 0.44 - 0.60, acetabulum: 0.36 - 0.76). In general, defects were grossly underestimated preoperatively. The level of experience did not influence the validity of measurements.
Our results indicate that these classifications do not provide valid and reliable assessment of femoral and acetabular bone stock loss prior to revision total hip arthroplasty.
本前瞻性研究旨在分析三种不同的放射学分类系统在评估翻修全髋关节置换术前骨量丢失情况时的有效性和可靠性。
采用以下分类系统进行研究：1）帕普罗斯基等人的分类系统；2）亨格福德等人的分类系统；3）德国骨科学与创伤外科学会（DGOT）的分类系统。连续纳入28例诊断为全髋关节置换术无菌性失败的患者。由三名临床经验水平不同的研究者采用盲法观察技术在术前对分类进行评估，分析评估者间的一致性。通过比较术前放射学检查结果（n = 3）和术中可视化及触诊对照结果（n = 1）分析有效性。使用斯皮尔曼相关系数（r(s)）确定多个有序变量之间的一致程度。
三名研究者之间进行配对比较的评估者间可靠性测试显示，相关系数不均匀（r(s)：帕普罗斯基股骨：0.45 - 0.67，髋臼：0.38 - 0.63；亨格福德：0.46 - 0.66；DGOT股骨：0.38 - 0.59，髋臼：0.42 - 0.76）。术前和术中检查结果之间的配对相关性分析再次显示相关系数不均匀（r(s)：帕普罗斯基股骨：0.59 - 0.68，髋臼：0.39 - 0.70；亨格福德：0.39 - 0.74；DGOT股骨：0.44 - 0.60，髋臼：0.36 - 0.76）。总体而言，术前对骨缺损严重低估。经验水平并未影响测量的有效性。
我们的结果表明，这些分类系统在翻修全髋关节置换术前对股骨和髋臼骨量丢失情况不能提供有效且可靠的评估。
