eGFR slope as a surrogate endpoint for end-stage kidney disease in patients with diabetes and eGFR > 30 mL/min/1.73 m2 in the J-DREAMS cohort

Abstract

Background: An analysis of European and American individuals revealed that a reduction in estimated glomerular filtration rate (eGFR) slope by 0.5 to 1.0 mL/min/1.73 m² per year is a surrogate endpoint for end-stage kidney disease (ESKD) in patients with early chronic kidney disease. However, it remains unclear whether this can be extrapolated to Japanese patients.

Methods: Using data from the Japan diabetes comprehensive database project based on an advanced electronic medical record system (J-DREAMS) cohort of 51,483 Japanese patients with diabetes and a baseline eGFR ≥ 30 mL/min/1.73 m², we examined whether the eGFR slope could be a surrogate indicator for ESKD. The eGFR slope was calculated at 1, 2, and 3 years, and the relationship between each eGFR slope and ESKD risk was estimated using a Cox proportional hazards model to obtain adjusted hazard ratios (aHRs).

Results: Slower eGFR decline by 0.75 mL/min/1.73 m²/year reduction in 1-, 2-, and 3-year slopes was associated with lower risk of ESKD (aHR 0.93 (95% confidence interval (CI) 0.92-0.95), 0.84 (95% CI 0.82-0.86), and 0.77 (95% CI 0.73-0.82), respectively); this relationship became more apparent as the slope calculation period increased. Similar results were obtained in subgroup analyses divided by baseline eGFR or baseline urine albumin-creatinine ratio (UACR), with a stronger correlation with ESKD in the baseline eGFR < 60 mL/min/1.73 m² group and in the baseline UACR < 30 mg/gCre group.

Conclusion: We found that changes in the eGFR slope were associated with ESKD risk in this population.

Keywords: Chronic kidney disease; Diabetic kidney disease; Surrogate endpoint; eGFR slope.

Fig. 1

Estimated glomerular filtration rate (eGFR) slope calculation and observation periods for events. The eGFR value at the beginning of the observation period was used as the baseline value. The 1-year eGFR slope was calculated for cases with eGFR records at two time points: baseline and 1 year ± 3 months from baseline. In the same way, 2- and 3-year eGFR slopes were calculated. The period from baseline until the date when the end-stage kidney disease event occurred or the last observation date was defined as the observation period for events (grey arrows), excluding the period used for the slope calculation (white arrows)

Fig. 2

Population distribution of change in estimated glomerular filtration rate (eGFR) slope and the association between eGFR slope and composite end-stage kidney disease (ESKD) events. The upper panel shows the spline curve of the association between composite ESKD events and the (A) 1-year, (B) 2-year, and (C) 3-year eGFR slopes. The lower panel shows the distribution of the number of cases in whom (A) 1-year, (B) 2-year, and (C) 3-year eGFR slopes were calculated. The green line corresponds to the slope calculated under the mixed-effects model, and the blue line corresponds to the slope calculated under ordinary least-squares methods. To plot the spline curve, the average value of the eGFR slope was used as a reference, and the number of knots was set to three

Fig. 3

Association between end-stage kidney disease risk and estimated glomerular filtration rate (eGFR) slope calculated under mixed-effects (ME) model or ordinary least-squares linear regression (OLS) model. (A) Composite ESKD risk or (B) pure ESKD risk associations with the 1-year, 2-year, and 3-year eGFR slope reduction of 0.75 mL/min/1.73 m²/year, calculated under mixed-effects (ME) model or ordinary least-squares linear regression (OLS) model. CI confidence interval; HR hazard ratio

Fig. 4

Association between composite end-stage kidney disease (ESKD) risk and estimated glomerular filtration rate (eGFR) slope: subgroup analysis. Composite ESKD risk associations with eGFR slope reduction of 0.75 mL/min/1.73 m²/year calculated under a mixed-effects model. eGFR slope was calculated over a period of (A) 1 year, (B) 2 years, and (C) 3 years, respectively. HRs and 95% CIs for each subgroup divided by eGFR (G1-2: ≥ 60 mL/min/1.73 m², G3: 30–60 mL/min/1.73 m²), or urine albumin-creatinine ratio (A1: < 30 mg/gCre, A2-3, ≥ 30 mg/gCre) were shown. CI confidence interval, HR hazard ratio.