P0786CLINICAL SIGNIFICANCE AND RELATED FACTORS OF GFR SLOPE IN A LARGE MULTICENTER OBSERVATIONAL STUDY IN JAPAN

Abstract

Abstract Background and Aims Recently, glomerular filtration rate (GFR) slope has attracted attention as an important surrogate marker for the prognosis of chronic kidney disease (CKD), with a reduction in slope of eGFR decline by 0.75 mL/min/1.73 m2 per year reportedly having clinical significance. As few large clinical studies on Japanese CKD patients exist, this investigation addresses the clinical significance of GFR slope and its related factors. Method To evaluate the clinical impact of GFR slope, we conducted a prognostic investigation of CKD patients in Japan by means of a large, multicenter, retrospective, observational study. Patients with CKD who were seen at among 15 general hospitals between January and March 2014 were surveyed using medical records. The selection criteria were age ≥20 years, estimated GFR (eGFR) <60 mL/min/1.73 m2, and receiving medical treatment for CKD. Baseline patient characteristics, eGFR changes, and hard endpoints (death or end-stage kidney disease requiring renal replacement therapy) during observation were analysed. We calculated GFR slope using GFR data of 2 years following the observation start point by 2 calculation methods, the linear mixed model and least squares linear regression, and examined the relationship of GFR slope with the hazard ratio of the composite hard endpoints. The factors related to GFR slope were also assessed by multiple regression analysis. Results Among a total of 11233 collected patients, we analyzed the data of 7490 CKD G3 and G4 patients whose GFR data during 2 years could be obtained (60% male, mean age: 71 years, CKD G3a: 55%, G3b: 30%, G4: 15%, mean eGFR: 44.1 mL/min/1.73 m2, urine protein positive: 51%, diabetes mellitus: 49%, use of RAS inhibitors: 57%). The mean observation period was 1040 days. Hard endpoints after the GFR slope measurement period occurred in 301 subjects. The GFR slope of the cohort was -0.948 mL/min/1.73 m2 per year (95% confidence interval [CI] -1.016, -0.880) in the linear mixed model and -0.982 mL/min/1.73 m2 per year (95% CI -1.075, -0.889) according to least squares linear regression. Both calculated GFR slopes were significantly related to the hazard ratio of the composite hard endpoints. Hazard ratio decreased by 0.85 (linear mixed model) and 0.9 (least squares linear regression) times in case of a reduction in slope of eGFR decline by 0.75 mL/min/1.73 m2 per year. Multiple regression analysis revealed strongly significant associations for GFR slope with urine protein and CKD stage and undetectable relationships for GFR slope with diabetes and age. Conclusion This study demonstrated the clinical significance of GFR slope as a surrogate marker for renal prognosis in Japanese CKD patients. In order to reduce slope of eGFR decline, active intervention for proteinuria before the progression to an advanced CKD stage appears to be effective.