Identification of the Joint Line in Revision Total Knee Arthroplasty Using a Multiple Linear Regression Model: a Cadaveric Study
Yunfei Hou,Jun Jiang,Han Liu,Ruikang Wang,Jingyu Wu,Yixiong Wang,Jianhao Lin
DOI: https://doi.org/10.1007/s00402-023-04792-3
IF: 2.928
2023-01-01
Archives of Orthopaedic and Trauma Surgery
Abstract:The results of revision total knee arthroplasty (rTKA) may be compromised by excessive joint line (JL) elevation. It is critical but challenging in reestablishing the JL in rTKA. Previous studies have confirmed that, biomechanically and clinically, JL elevation should not exceed 4 mm. Image-based studies described several approaches to locate the JL intraoperatively, however magnification errors could occur. In this cadaveric study, we aim to define an accurate and reliable method to determine the JL. Thirteen male and eleven female cadavers were used, with an average age of death being 48.3 years. The transepicondylar width (TEW), the distance from the medial (MEJL) and lateral (LEJL) epicondyle, adductor tubercle (ATJL), fibular head (FHJL) and tibial tubercle (TTJL) to the JL were measured in 48 knees. Intra- and interobserver reliability and validity were tested prior to any additional analysis. Pearson correlation and linear regression analysis were used to examine the correlations between landmark-JL distances (LEJL, MEJL, ATJL, FHJL and TTJL) and the TEW, and to further derive models for intraoperative JL determination. The accuracy of different models, quantified by errors between estimated and measured landmark-JL distances, was compared using the Friedman and post hoc Dunn tests. The intra- and inter-observer measurements for TEW, MEJL, LEJL, ATJL, TTJL and FHJL did not differ significantly (p > 0.05). Between genders, significant differences were found on TEW, MEJL, LEJL, ATJL, FHJL and TTJL (p < 0.05). There was no association between TEW and either FHJL or TTJL (p > 0.05), while ATJL, MEJL, and LEJL were found to be correlated with TEW (p < 0.05). Six models were derived: (1) MEJL = 0.37*TEW (r = 0.384), (2) LEJL = 0.28*TEW (r = 0.380), (3) ATJL = 0.47*TEW (r = 0.608), (4) MEJL = 0.413*TEW − 4.197 (R2 = 0.473), (5) LEJL = 0.236*TEW + 3.373 (R2 = 0.326), (6) ATJL = 0.455*TEW + 1.440 (R2 = 0.556). Errors were defined as deviations between estimated and actual landmark-JL distances. The mean absolute value of the errors, created by Model 1–6 was 3.18 ± 2.25, 2.53 ± 2.15, 2.64 ± 2.2, 1.85 ± 1.61, 1.60 ± 1.59 and 1.71 ± 1.5, respectively. The error could be limited to 4 mm in 72.9%, 83.3%, 72.9%, 87.5%, 87.5%, and 93.8% of the cases by referencing Model 1–6, respectively. Compared to previous image-based measurements, the current cadaveric study most closely resembles a realistic view of intraoperative settings and could circumvents magnification errors. We recommend using Model 6, the JL can be best estimated by referencing the AT and the ATJL can be calculated as ATJL (mm) = 0.455*TEW (mm) + 1.440 (mm).