Mylène Jansen

288 Chapter 14 Sensitivity analyses, correcting for the fact that patients in almost all comparisons were included in different trials, are shown in Supplementary Table S1. Correcting for the trial did not change significance for the primary outcome (change in MAC cartilage thickness) or for MAC JSW change. For the change in denuded bone area the difference was no longer statistically significantly different for any comparison when correcting for the original trial patients were included in. Discussion The main goal of this study was to compare 2-year quantitative cartilage changes during treatment with KJD versus HTO, hypothesizing that KJD is more effective in restoring cartilage in the MAC, while avoiding negative effects on the LAC. The secondary goal was to identify factors that can predict cartilage restoration activity. Baseline OA severity was the strongest indicator of cartilage restoration response after treatment, independent of treatment and only severe knee OA patients showed statistically significant cartilage restoration after treatment in both cartilage thickness and denuded bone area, in accordance with radiographic JSW results used as reference. Contrarily, HTO treated patients showed statistically significant cartilage loss on MRI, while the radiographic JSW of the MAC increased. Patients that received KJD in case of HTO indication, with relatively mild OA as compared to KJD in case of TKA indication, demonstrated no differences in cartilage restoration when compared to HTO treated patients. Effect sizes were moderate to large and the changes, although seemingly small in mm and percentage of area, seem to be clinically significant as compared to natural progression of loss in cartilage thickness and increase in denuded bone area in comparable untreated knee OA patients. 19 Discussion of the subregional results can be found in the supplementary data. 36,37 The leg axis deviation, the main reason to indicate a knee OA patient for treatment with HTO, did not have a significant influence on the amount of cartilage restoration (supplementary data). Instead, along with the Kellgren-Lawrence grade, a higher patient age and lower baseline cartilage thickness were the strongest indicators for greater cartilage restoration, likely because both these parameters are significantly associated with a higher Kellgren-Lawrence grade (1-way ANOVA: both p< 0.045; data not shown). A Kellgren-Lawrence grading providing compartment-specific instead of knee-specific scores was applied and there was only 1 KJD patient whose LAC was scored with a Kellgren-Lawrence grade >2. This patient displayed a relatively large increase in cartilage thickness in the LAC that was comparable to that in the MAC (MAC +0.22m; LAC +0.24 mm). Treatment with HTO demonstrated both cartilage thickness loss and increase in denuded bone area by MRI of the MAC whereas radiographic MAC JSW increased. It is therefore likely that the MAC JSW increase is predominantly a result of pseudo-widening, due to a mechanical