Figure 4

Time-course changes in estimated glomerular filtration rates (eGFRs) from week 0 through week 108 of treatment in subcohorts according to cross-classification. (a) Quadrant 1: without proteinuria and a serum creatinine level ≥ median. (b) Quadrant 2: without proteinuria and a serum creatinine level < median. (c) Quadrant 3: with proteinuria and a serum creatinine level ≥ median. (d) Quadrant 4: with proteinuria and a serum creatinine level < median. The mean eGFR in the two groups is shown at different timepoints during the trial. Error bars indicate the standard error. The mean eGFR is shown for participants with available levels at each timepoint. P-values were obtained by a test of the trend profile using a mixed model. Similar to Fig. 2, quadrants 1 and 4 can be used to discriminate the importance of the absence of proteinuria and a serum creatinine level < median for the efficacy of febuxostat in patients with chronic kidney disease. As the trend profile in the time-course changes in the eGFRs in quadrant 1 was significant for the efficacy of febuxostat, despite the fact that the serum creatinine level was ≥ median, the absence of proteinuria is a more important attribute compared to the serum creatinine level being < median. Accordingly, the absence of proteinuria is an attribute affecting the efficacy of febuxostat in patients with chronic kidney disease.