Creatinine- vs. cystatin C-based equations compared with 99mTcDTPA scintigraphy to assess glomerular filtration rate in chronic kidney disease.

BACKGROUND: In chronic kidney disease (CKD), accurate estimation of the glomerular filtration rate (GFR) is mandatory. Gold standard methods for its estimation are expensive and time-consuming. We compared creatinine- versus cystatin C-based equations to measure GFR, employing (99m)Tc-DTPA scintigraphy as the gold standard.

METHODS: This was a prospective cross-sectional observational study including 300 subjects. CKD was defined according to K/DOQI guidelines, and patients were separated into groups: stage 1 (G1), n=26; stage 2 (G2), n=52; stage 3 (G3), n=90; stage 4 (G4), n=37; stage 5 (G5), n=60; and control group, n=35. Creatinine-based estimates were from 24-hour creatinine clearance using the Walser formula, Cockcroft-Gault, MDRD-4 and CKD-EPI; cystatin C equations used were Larsson, Larsson modified equation, Grubb and Hoek.

RESULTS: Age and body mass index were different among groups; proteinuria, hypertension, diabetes and primary glomerulopathies significantly increased as CKD worsened. In the global assessment, CKD-EPI and Hoek gave the highest correlations with (99m)Tc-DTPA: rho=0.826, p<0.001 and rho=0.704, p<0.001, respectively. Most significant linear regressions obtained: CKD-EPI vs. (99m)Tc-DTPA, Hoek vs. (99m)Tc-DTPA and CKD-EPI vs. Hoek. However, important differences emerged when each group was analyzed separately. Best significant correlations obtained with (99m)Tc-DTPA: control group, creatinine clearance rho=0.421, p=0.012; G1, Crockoft-Gault rho=0.588, p=0.003; G2, CKD-EPI rho=0.462, p<0.05; G3, CKD-EPI rho=0.508, p<0.001; G4, Hoek rho=0.618, p<0.001; G5, CKD-EPI rho=0.604, p<0.001.

CONCLUSIONS: At GFR <60 ml/min, CKD-EPI and Hoek equations appeared to best correlate with (99m)TcDTPA. In controls and at early stages of CKD, creatinine-based equations correlated better with (99m)Tc-DTPA, with CKD-EPI being the one with the best degree of agreement.