Mylène Jansen

Long-term clinical success of KJD 49 3 Statistical analysis Baseline characteristics and initial clinical and structural improvement (first year post- treatment) were compared in survivors and patients that underwent TKA during follow-up. For the baseline characteristics, normally distributed data are described with the mean and SD while not normally distributed and categorical data are described with the median and interquartile range (IQR). For comparing the characteristics between groups, independent t -tests were used for normally distributed data, Mann-Whitney U Tests were used when baseline data were not normally distributed in either or both groups, and the chi-square test was used for categorical variables. Survival analysis was performed using the Kaplan-Meier estimator, where event occurrence was defined by patients undergoing TKA. Patients were censored when withdrawing consent for further follow-up or after the maximum follow-up period of 9 years. Survival was compared for all baseline characteristics and initial clinical and structural change using the log-rank test, for which continuous data was divided into groups based on the distribution of the variable. For JSW measures, the smallest detectable difference was used as cut-off. 11 To evaluate whether clinical and structural follow-up values significantly differed from baseline values, 2-sided paired t -tests were used. In case of not normally distributed baseline or follow- up data, the Wilcoxon Signed-Rank Test was used instead. The tests were performed separately for survivors and those who got TKA over time. Since failures have no follow-up results after receiving TKA, last-measured outcomes were used for comparison with baseline values. Nine- year clinical and structural outcomes for survivors were compared with last-reported outcomes for patients receiving TKA using independent t -tests. Mean changes and 95% confidence interval (95%CI) are given for clinical outcome parameters, mean and standard deviation (SD) before and after treatment are given for structural parameters, and median and interquartile range (IQR) before and after treatment for not normally distributed data. Long-term survival of KJD was predicted using Cox regression analyses, where the influence of different covariates on survival time was analyzed. Covariates were all baseline characteristics and initial clinical and structural improvement separately. Initial improvement as potential predictors were corrected for baseline values by adding them as covariates. Survival prediction effects are estimated with a hazard ratio (HR) with 95%CI. P- values <0.05 were considered statistically significant. For all statistical tests, IBM SPSS Statistics version 20.0.0 was used. Results Patients During 9 years of follow-up, 3 patients withdrew their consent for further follow-up. Nine patients underwent TKA after on average 6.4 years (range 3.8–9.0 years), leaving 8 survivors