Mylène Jansen

334 Chapter 16 For example, for the effect of change in FGF-2 over 6-weeks, on change in KOOS 4 over 6-months, for a 1-unit increase in FGF-2 change per pg/ml, the change in KOOS 4 is 0.03 points. To interpret the 95%CI, the underlying effect in the population could lie between 0.004 and 0.057. To aid interpretation, the median increase of FGF-2 over 6 weeks is 165 pg/ml. Hence for a 165 pg/ml unit increase in FGF-2, the change in KOOS4 is 4.95 points (95%CI 0.66–9.41). IL-8 had the largest and increasing effect size (0.28 by 12 months), but the confidence intervals at all time points were wide. Figure 3 : Correlation between change of analytes in the synovial fluid of participants over period of knee joint distraction. Spearman rank tests were performed to determine correlations between the change in levels of synovial fluid analytes over the 6 week distraction period (concentration at 6 weeks - baseline concentrations). Correlation coefficients were calculated using all available participant data and the mean of 2 repeated (duplicate) measures for each synovial fluid sample. Strength of correlation by Spearman R coefficient is shown: * (Mid gray shading): Low positive (negative) correlation, 0.30–0.49 (-0.30 to -0.49). ** (Dark gray shading): Moderate positive (negative) correlation, 0.50–0.69 (-0.50 to -0.69). FGF-2: basic fibroblast growth factor; IL-6: interleukin 6; IL-8: interleukin 8; LTBP2: latent-transforming growth factor beta-binding protein 2; MCP-1: monocyte chemoattractant protein 1; MMP3: matrix metalloproteinase 3; TGF β -1: transforming growth factor beta 1; TIMP-1: tissue inhibitor of metalloproteinases 1; TSG-6: tumor necrosis factor-inducible gene 6 protein. To test the relevance of these findings, we categorized participants’ molecular measurements as having no relevant change, a relevant increase or a relevant decrease over the 6 week distraction period (see methods) and examined the association of these categories with change in KOOS 4 . Those with a relevant increase in SF IL-8 during the distraction period had a greater improvement in KOOS 4 over 12 months than those with no change (regression coefficient 17.6 (1.2–34.0); p= 0.04). However, no other molecular changes were associated with clinical outcome when categorized in this way (Supplementary Table 3). Furthermore, the confidence intervals for this observation are wide and given that the other findings for IL-8 did not reach significance (Figure 4A), this could be a chance finding).