Aernoud Fiolet

237 Short-term effect of low-dose colchicine on biomarkers a central core laboratory. Hs-CRP was measured using a research hs-CRP Elisa kit (Hycult Biotech #HK369, Uden, the Netherlands). The lower detection limit of this assay was 0.4ng/L and the inter- and intra-assay coefficients of variation were <6.9% and <6.3% respectively. IL-6 levels were measured by highly sensitive human IL-6 immunoassay (R&D Systems #D6050, Minneapolis, MN, USA) for which intra-assay and inter-assay coefficients of variation ranged from 4.2% to 6.4%. This assay had a sensitivity of 0.7 pg/mL. The hematology parameters were determined with the XN9000 of Sysmex (Sysmex, Kobe, Japan). The serum LDL cholesterol was determined with the third- generation homogeneous enzymatic colorimetric assay, the serum creatinine with the second-generation enzymatic assay. These analyses were carried out in a routine setting under ISO15189 compliance. Statistical analysis We estimated that using a sample size of 130 subjects would provide 80% power to detect a mean of the differences of -1 mg/dl in hs-CRP concentration, assuming a standard deviation of 4 mg/dl. Due to the expected non-normal distribution of differences, the minimal number of participants was extendedwith 15% to increase discriminative power in non-parametric testing. 20 Central tendencies and distribution of continuous parameters were displayed using mean or median with standard deviation or 25th and 75th percentile in case of normally and non-normally distributed variables respectively. Categorical variables were presented as proportions. Paired differences of parameters were evaluated and provided using the mean or median of the differences and the corresponding 95% confidence interval (CI). Standard errors and confidence intervals for medians in non-parametric distributions were computed using bias- corrected and accelerated bootstrapping. A Hodges–Lehmann estimator was used to provide a pseudo-median and confidence interval of the differences between non-parametric distributions. Formal hypothesis testing was done using a paired sample T–test for normally distributed differences and the Wilcoxon Signed Rank Test with continuity correction for non-normally distributed differences.The linear relationship of two continuous parameters was calculated using the Spearman's rank correlation coefficient.