Background: There is increased interest in surrogate endpoints for clinical trials of chronic kidney disease.
Methods: In this nationwide observational study of 456 patients with type 2 diabetes and clinically suspected diabetic nephropathy followed for a median of 4.2 years, we evaluated the association between estimated glomerular filtration rate (eGFR) and albuminuria at baseline or during follow-up and risk of ESRD.
Results: Low eGFR (<60 mL/min/1.73 m2) and macroalbuminuria at enrollment were independently associated with risk of ESRD. In patients with macroalbuminuria, both ≤-50% change and -50 to -30% change in eGFR over 1 and 2 years were predictive of ESRD. The higher cut point (≥50% decline in eGFR) was more strongly predictive but less common. Remission of macroalbuminuria to normo-/microalbuminuria at 1 and 2 years was associated with a lower incidence of ESRD than no remission; however, it was not a determinant for ESRD independently of initial eGFR and initial protein-to-creatinine ratio.
Conclusion: These results suggest that a ≥30% decline in eGFR over 1 or 2 years adds prognostic information about risk for ESRD in patients with type 2 diabetes and macroalbuminuria, supporting the consideration of percentage decline in eGFR as a surrogate endpoint among macroalbuminuric cases in type 2 diabetes. On the other hand, our study suggests that additional analyses on the relationship between remission of macroalbuminuria and risk of ESRD are needed in type 2 diabetes.
Keywords: Changes in estimated glomerular filtration rate; Diabetic nephropathy; End-stage renal disease; Remission of macroalbuminuria; Surrogate endpoints.