Comparison of estimated GFR using cystatin C versus creatinine in pediatric kidney transplant recipients

Abstract

Abstract Background An accurate, rapid estimate of glomerular filtration rate (GFR) in kidney transplant patients affords early detection of transplant deterioration and timely intervention. This study compared the performance of serum creatinine (Cr) and cystatin C (CysC)-based GFR equations to iohexol GFR (iGFR) among pediatric kidney transplant recipients. Methods CysC, Cr, and iGFR were obtained from 45 kidney transplant patients, 1–18 years old. Cr- and CysC-estimated GFR (eGFR) was compared against iGFR using the Cr-based (Bedside Schwartz, U25-Cr), CysC-based (Gentian CysC, CAPA, U25-CysC), and Cr-CysC combination (CKiD Cr-CysC, U25 Cr-CysC) equations in terms of bias, precision, and accuracy. Bland-Altman plots assessed the agreement between eGFR and iGFR. Secondary analyses evaluated the formulas in patients with biopsy-proven histological changes, and K/DOQI CKD staging. Results U25-CysC and Gentian CysC equations had the smallest bias. 88.9% of U25-CysC and 82.2% of Gentian CysC estimations were within 30% iGFR; 37.8% of both and 40% of CKiD Cr-CysC were within 10% iGFR. In subjects with histological changes on biopsy, U25-CysC and Gentian CysC had the smallest bias and were most accurate - both with 83.3% of and 41.7% of estimates within 30% and 10% iGFR, respectively. U25-CysC, CKiD Cr-CysC, and U25 Cr-CysC, were the most precise. Bland-Altman plots show the Bedside Schwartz, Gentian CysC, CAPA, and U25-CysC tend to overestimate GFR when > 100 ml/min/1.72m2. CAPA misclassified CKD stage the least (whole cohort 24.4%, histological changes on biopsy 33.3%). Conclusion In this small cohort, CysC-based equations appear to have better bias, precision, and accuracy in predicting GFR.