Abstract Background and Aims In recent years, growing evidence has shown the prognostic utility of eGFR slope for the risk of end-staging renal disease (ESRD). Previous observational studies have assessed the association between renal events and eGFR slopes estimated by two measurements of eGFR but had great differences in the interval between two eGFR measurements. In this study, we thus aimed to determine the appropriate interval between two eGFR measurements to evaluate the association of eGFR slope with incidence of renal events. Method This is a retrospective cohort study in 203 chronic kidney disease (CKD) patients who visited Nara Prefecture General Medical Center, Japan between 1 January 2013 and 31 December 2016 and in whom two or more than two measurements of eGFR levels were confirmed in medical records. eGFR slopes were estimated by using two measurements of eGFR at baseline and 0.5, 1, 1.5, 2, or 3 years. We excluded patients with acute kidney injury, urologic malignancies, nephrotic syndrome with steroid treatment or collagen diseases. Outcome was renal events defined as a composite of ESRD and eGFR decline of > 30%. C statistics were used to evaluate the association between eGFR slope and incidence of renal events. Results The median (interquartile range) age of study participants was 67 (56-77) years and 71 (37%) were male. The median (interquartile range) levels of baseline eGFR were 34 (21-48) mL/min/1.73m2 and diabetes was present in 80 (39%) participants. During the median follow-up period of 38 months, renal events occurred in 52 participants. Medial levels of eGFR slopes0.5, 1, 1,5, 2 and 3yr were -7.8, -3.6, -2.9, -0.9, -1.5 mL/min/1.73m2, respectively. C-statistics of eGFR slopes0.5, 1, 1,5, 2 and 3yr for renal events were 0.622, 0.691, 0.797, 0.858, 0.806, respectively (shown in Figure 1), and that of eGFR slope1.5yr was significantly higher than that of eGFR slope0.5yr and eGFR slope1yr (p<0.001 and p=0.001, respectively). In stratified analysis, eGFR slope1.5yr had higher prognostic ability of renal events in patients with versus without diabetes, advanced CKD and proteinuria. C-statistics of renal events when considering baseline eGFR alone was 0.853 but combination use of baseline eGFR and eGFR slope1.5yr significantly increased c-statistics, to 0.913 (p=0.01). Conclusion eGFR slope for high prognostic ability of renal events may be needed to be calculated by at least 1.5-year interval between two eGFR measurements.
Read full abstract