A comparison between serum creatinine and cystatin C-based equations for estimation of graft function

Background: Accurate monitoring of estimated glomerular filtration rate (GFR) is essential for an optimal management of kidney transplant (KT) patients. We aimed to compare the predictive performance of creatinine- and cystatin C-based GFR with creatinine clearance (CCr) in a 24-hour urine collection as the reference test.

Methods: GFR was calculated using cystatin C-based equations (Le Bricon, Stevens) and two creatinine-based equations [Cockcroft-Gault (CG), modification of diet in renal disease (MDRD)]. We enrolled 173 KT recipients. Bias, precision, and accuracy of each equation were determined. Kappa statistics evaluated the concordance between the reference test and GFR formulas in classifying patients according graft function (CCr <60 or ≥60 mL/min/1.73 m2).

Results: Patients (108 males) had a mean age of 48.6 ± 12.2 years and a median posttransplant time of 6.8 years. Mean CCr was 69.3 ± 19.9 (range: 32.1-105.2) mL/min/1.73 m2. The cystatin C-based equations estimates (Le Bricon, Stevens) had the highest accuracy (83.8% and 87.9% within 30% of CCr result, respectively). Precision of Le Bricon, Stevens, and MDRD was similar (around 13.5 mL/min/1.73 m2)) and much better than CG (22.5 mL/min/1.73 m2). The lowest bias was seen in Le Bricon (-1.2 mL/min/1.73 m2), followed by CG, Stevens, and MDRD (-2.6, -9.5, -16.5 mL/min/1.73 m(2), respectively). Kappa coefficient was higher in cystatin C-based equations (0.53) in contrast with CG (0.48) and MDRD (0.40). Stevens had a high sensitivity (90.8%) and low specificity (66.7%) and, conversely, Le Bricon had 64.6% sensitivity and 87.0% specificity.

Conclusions: Cystatin C-based equations showed a better predictive performance of graft function than creatinine-based equations. The role of cystatin C in graft function monitoring deserves further investigation.