Comparing GFR Estimating Equations Using Cystatin C and Creatinine in Elderly Individuals

• Published in: Journal of the American Society of Nephrology Vol. 26; no. 8; pp. 1982 - 1989
• Main Authors: Fan, Li, Levey, Andrew S, Gudnason, Vilmundur, Eiriksdottir, Gudny, Andresdottir, Margret B, Gudmundsdottir, Hrefna, Indridason, Olafur S, Palsson, Runolfur, Mitchell, Gary, Inker, Lesley A
• Format: Journal Article
• Language: English
• Published: United States American Society of Nephrology 01.08.2015
• Subjects: Aged; Aged, 80 and over; Algorithms; Clinical Epidemiology; Creatinine - blood; Cystatin C - blood; Female; Glomerular Filtration Rate; Humans; Male; cystatin C; GFR; creatinine; CKD; elderly
• ISSN: 1046-6673; 1533-3450
• DOI: 10.1681/ASN.2014060607

Current guidelines recommend reporting eGFR using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equations unless other equations are more accurate, and recommend the combination of creatinine and cystatin C (eGFRcr-cys) as more accurate than either eGFRcr or eGFRcys alone. However, preferred equations and filtration markers in elderly individuals are debated. In 805 adults enrolled in the community-based Age, Gene/Environment Susceptibility (AGES)-Reykjavik Study, we measured GFR (mGFR) using plasma clearance of iohexol, standardized creatinine and cystatin C, and eGFR using the CKD-EPI, Japanese, Berlin Initiative Study (BIS), and Caucasian and Asian pediatric and adult subjects (CAPA) equations. We evaluated equation performance using bias, precision, and two measures of accuracy. We first compared the Japanese, BIS, and CAPA equations with the CKD-EPI equations to determine the preferred equations, and then compared eGFRcr and eGFRcys with eGFRcr-cys using the preferred equations. Mean (SD) age was 80.3 (4.0) years. Median (25th, 75th) mGFR was 64 (52, 73) ml/min per 1.73 m(2), and the prevalence of decreased GFR was 39% (95% confidence interval, 35.8 to 42.5). Among 24 comparisons with the other equations, CKD-EPI equations performed better in 9, similar in 13, and worse in 2. Using the CKD-EPI equations, eGFRcr-cys performed better than eGFRcr in four metrics, better than eGFRcys in two metrics, and similar to eGFRcys in two metrics. In conclusion, neither the Japanese, BIS, nor CAPA equations were superior to the CKD-EPI equations in this cohort of community-dwelling elderly individuals. Using the CKD-EPI equations, eGFRcr-cys performed better than eGFRcr or eGFRcys.