Discordant Results Between Creatinine- and Cystatin C-based Equations for Estimating GFR

Discordant results between cystatin C and creatinine in estimating glomerular filtration rate (GFR) are an important medical issue. However, the equation that should be used when GFR estimates are discordant remains unclear.

This cross-sectional analysis included 15,485 participants with GFR measured by the clearance of an exogenous marker, serum creatinine, and cystatin C. We studied the proportion of discordant results defined as an absolute (> 15 ml/min per 1.73 m²) or relative (> 20%) difference between creatinine-based estimated GFR (eGFR, eGFR_crea) and cystatin C-based eGFR (eGFR_cys) using different equations (Chronic Kidney Disease Epidemiology Collaboration [CKD-EPI], European Kidney Function Consortium [EKFC], and reexpressed Lund-Malmö [r-LMR]). We also researched for the best estimating equations to be used in subjects with concordant or discordant results to estimate measured GFR (mGFR).

In the total cohort, the proportion of subjects with discordant results (absolute or relative) was larger for CKD-EPI (35.1 and 40.6%) than for EKFC (21.9 and 31.7%) or r-LMR (22.8 and 32.8%) equations. Among discrepant results, the proportion of eGFR_cys < eGFR_crea was significantly higher than the proportion of eGFR_crea < eGFR_cys for the CKD-EPI equations, whereas the occurrence of discrepancy was similar in the 2 discrepant groups for EKFC or r-LMR. For the EKFC and r-LMR equations, but not for the CKD-EPI, the equation combining creatinine and cystatin C was consistently the closest to the mGFR in the discrepant groups.

Based on the lower discrepancy proportion, better balance between eGFR_crea and eGFR_cys, and better concordance with mGFR, the EKFC, and r-LMR equations should be preferred over the CKD-EPI to estimate GFR.