Are cystatin C-based equations superior to creatinine-based equations for estimating GFR in Chinese elderly population?

Cystatin C has been proposed as a surrogate marker of kidney function. The elderly population accounts for the largest proportion of chronic kidney disease (CKD) patients. The aim of this study was to assess the diagnostic value of serum cystatin C and compare the applicability of cystatin C-based equations with serum creatinine (Scr)-based equations for estimating glomerular filtration rate (GFR).

The estimated GFR (eGFR) values from six cystatin C-based equations (Tan, MacIsaac, Ma, Stevens1-3) and three Scr-based equations (CG, MDRD, CKD-EPI) were compared with the reference GFR (rGFR) values from ^99mTc-DTPA renal dynamic imaging method.

A total of 110 elderly Chinese (60–92 year, 71.05 ± 7.62 year) were enrolled. Cystatin C had better diagnostic value than Scr (relationship coefficient with rGFR: cystatin C −0.847 vs. Scr −0.729, P < 0.01; sensitivity: cystatin C 0.90 vs. Scr 0.55, P < 0.01; AUC^ROC: cystatin C 0.857 vs. Scr 0.757, P < 0.01). All the equations predicted GFR more accurately for rGFR ≥ 60 ml/min/1.73 m² than for rGFR < 60 ml/min/1.73 m². Most equations had acceptable accuracy. The cystatin C-based equations deviated from rGFR by −12.78 ml/min/1.73 m² to −2.12 ml/min/1.73 m², with accuracy varying from 64.6 to 82.7 %. The Scr-based equations deviated from rGFR by −5.37 ml/min/1.73 m² to −0.68 ml/min/1.73 m², with accuracy varying from 77.3 to 79.1 %. The CKD-EPI, MacIsaac and Ma equations predicted no bias with rGFR (P > 0.05), with higher accuracy and lower deviation in the total group. The MacIsaac, CKD-EPI and Stevens3 equations could be optimal for those with normal and mildly impaired kidney function, whereas the Ma equation for those with CKD.

Cystatin C is a promising kidney function marker. However, not all cystatin C-based equations could be superior to the Scr-equations.