Mylène Jansen

MRI cartilage thickness after KJD and HTO 297 14 Location-independent results The location-independent cartilage thickening scores were 0.81mm (SD 0.93) for KJD HTO , 0.55 (0.48) for HTO, and 1.62 (0.95) for KJD TKA . The thinning scores were -1.14 (0.95) for KJD HTO , -1.14 (1.48) for HTO, and -0.72 (0.69) for KJD TKA . Kruskal-Wallis tests showed no statistically significant difference between the 3 groups in thinning scores ( p= 0.23) but the thickening score was significantly greater for KJD TKA than for the other 2 groups as confirmed by post-hoc Dunn tests (KJD HTO p= 0.016; HTO p= 0.001). Yet, this was no longer true for the comparison KJD TKA versus KJD HTO when correcting for significantly different baseline characteristics using linear regression ( p= 0.505). Prediction of cartilage thickness changes A multivariable linear regression model, using patient baseline characteristics and baseline MAC cartilage thickness as independent variables, revealed that only the Kellgren-Lawrence grade was a significant predictor ( B =0.105; p= 0.01) of MAC cartilage thickness change in all KJD and HTO patients together: a higher Kellgren-Lawrence grade was associated with in a greater increase in cartilage thickness during treatment. Using treatment as independent variable resulted in a better fit of the univariable model ( R 2 =0.120 with p= 0.004 compared to R 2 =0.095 with p= 0.01), therefore KJD and HTO treated patients were evaluated in separate models as well. For HTO patients, none of the parameters, including leg axis deviation, significantly predicted the MAC cartilage thickness or JSW change. In contrast, in KJD patients a multivariable linear regression model left only Kellgren-Lawrence grade as a significant predictor for MAC cartilage thickness change. Univariable models showed that patient age ( B =0.018; R 2 =0.128; p= 0.04), baseline MAC cartilage thickness ( B =-0.165; R 2 =0.207; p= 0.006) and patient group KJD TKA /KJD HTO ( B =- 0.245; R 2 =0.133; p= 0.03) significantly predicted cartilage change as well, but Kellgren- Lawrence grade ( B =0.174; R 2 =0.255; p= 0.002) remained the strongest predictor.