Agreement between cystatin-C and creatinine based estimated glomerular filtration rate among Ethiopian children.

IntroductionGlomerular filtration rate (GFR) is invasive to measure. Therefore, in clinical care, estimated GFR is derived from serum levels of endogenous filtration markers such as creatinine and cystatin C. Multiple studies from high income countries showed differences between estimated glomerular filtration rate based on cystatin C (eGFRcys) and creatinine (eGFRcr). This study aimed to assess the agreement between eGFRcys and eGFRcr in Ethiopian children and identify factors influencing higher eGFRcys and eGFRcr. MethodWe studied 350 Ethiopian children who were part of the iABC birth cohort study. At the recent follow-up (average age 10 years), serum cystatin C and creatinine were measured. Formulas by Berg (2015) and Hoste (2014) were used to estimate eGFRcys and eGFRcr, respectively, and Bland-Altman plots assessed their agreement. The difference in eGFR (eGFRdiff) was calculated and categorized as less than-15 mL/min/1.73 m{superscript 2} (higher eGFRcr), between-15 and <15 mL/min/1.73 m{superscript 2} (concordant), and greater than or equal to 15 mL/min/1.73 m{superscript 2} (higher eGFRcys). Multinomial logistic regression was used to identify factors associated with higher eGFRcr and higher eGFRcys. ResultEstimated glomerular filtration rate (eGFR) showed significant variation based on the estimation formula used. When using formulas by Berg (2015) and Hoste (2014), the median (IQR) eGFRcys and eGFRcr were 99.4 (90.0; 114.1), and 123.2 (110.3; 143.8) mL/min/1.73 m2, respectively. Overall, we observed a poor agreement between eGFRcys and eGFRcr, with only 94 (27.6%) children having concordant results compared to 220 (64.7%) with higher eGFRcr and 26 (7.6%) with higher eGFRcys. If the eGFRcys results are considered reliable, 27.5% of the children had eGFR below 90 mL/min/1.73 m{superscript 2}. ConclusionThere was very marked variation in the distributions of estimated eGFRs depending on which formulas for children were used. Agreement between eGFR estimated using cystatin C and creatinine was poor among Ethiopian children. Relative to eGFRcys, kidney function may be overestimated by creatinine-based equation as up to 30ml/min in Ethiopia. Ideally, a validation study with GFR measured by gold standard methods (Inlulin clearance) among children is required. However, because of its invasive nature and financial concerns, Iohexol clearance studies are recommended.