Comparison of cystatin C and creatinine-based equations for GFR estimation after living kidney donation.

BACKGROUND: The performance of glomerular filtration rate (GFR) equations incorporating both cystatin C (CysC) and serum creatinine (Creat) in living kidney donors has not been studied before. METHODS: From a population of 3,698 living kidney donors, 257 donors were randomly selected to undergo GFR measurement (mGFR) by the plasma disappearance of iohexol. GFR was estimated with the Modification of Diet in Renal Disease (MDRD) equation and the Chronic Kidney Disease Epidemiology Collaboration study eGFR(CKD-EPI-Creat) in 257 donors and the two newly developed equations using CysC with and without Creat, eGFR(CKD-EPI-CysC) and eGFR(CKD-EPI-Creat+CysC), in 215 donors. RESULTS: Mean mGFR was 71.8±11.8 mL/min/1.73 m. The eGFR(MDRD) exhibited least and only negative bias and the three other models were comparable in terms of bias. The eGFR(CKD-EPI-Creat+CysC) equation was most precise; r=0.64. Both eGFR(MDRD) and eGFR(CKD-EPI-Creat+CysC) had high percentage (94.4% and 92.6%, respectively) of estimates falling within 30% of mGFR versus estimates by eGFR(CKD-EPI-Creat) and eGFR(CKD-EPI-CysC) equations (87.2% and 85.1%, respectively). The eGFR(MDRD) was by far most accurate in identifying those with mGFR less than 60 mL/min/1.73 m whereas the CKD-EPI models were extremely accurate in classifying those with mGFR greater than or equal to 60 mL/min/1.73 m. CONCLUSIONS: eGFR(CKD-EPI-Creat+CysC) equation provides comparable accuracy to the eGFR(MDRD) in overall estimation of mGFR, but with higher precision. However, eGFR(CKD-EPI-Creat+CysC) clearly misses many of those with a post-donation GFR less than 60 mL/min/1.73 m and therefore eGFR(MDRD) is preferable in detecting donors with GFR less than 60 mL/min/1.73 m.