Albuminuria and estimated glomerular filtration rate as predictors of diabetic end-stage renal disease and death.

BACKGROUND AND OBJECTIVES

We investigated predictive value of albuminuria and estimated GFR (eGFR) for ESRD in Pima Indians with type 2 diabetes.

DESIGN, SETTING, PARTICIPANTS AND MEASUREMENTS: Beginning in 1982, 2420 diabetic Pima Indians ≥18 years old were followed until they developed ESRD or died or until December 31, 2005. Individuals were classified at baseline by urinary albumin-to-creatinine ratio (ACR) and by eGFR, calculated by the Chronic Kidney Disease Epidemiology Collaboration equation. Predictors of ESRD and mortality were examined by proportional hazards regression.

RESULTS

During a mean follow-up of 10.2 years, 287 individuals developed ESRD. Incidence of ESRD among individuals with macroalbuminuria (ACR ≥ 300 mg/g) was 9.3 times that of those with normoalbuminuria (ACR < 30 mg/g), controlled for age, gender, and duration of diabetes. Incidence among individuals with eGFR 15 to 29 ml/min per 1.73 m(2) was 81.9 times that of those with eGFR 90 to 119 ml/min per 1.73 m(2). Models that combined albuminuria and eGFR added significant predictive information about risk of ESRD or death compared with models containing eGFR or albuminuria alone. The hazard ratio for ESRD associated with a 10-ml/min per 1.73 m(2) lower eGFR was 1.36, whereas that associated with an increase in albuminuria category was 2.69; corresponding hazard ratios for death were 1.15 and 1.37.

CONCLUSIONS

These results suggest that incorporation of quantitative information about albuminuria into staging systems based on eGFR adds significant prognostic information about risk for diabetic ESRD and death.