[Clinical application of full age spectrum formula based on serum creatinine in patients with chronic kidney disease].

OBJECTIVE: To compare the estimated glomerular filtration rates (eGFR) by five formulas based on serum creatinine (SCr), and to explore the diagnostic efficacy of full age spectrum (FAS) equation based on SCr for renal insufficiency in patients with chronic kidney disease (CKD).

METHODS: 2 219 patients with CKD admitted to department of nephrology of Guangdong Provincial Second People's Hospital from December 2015 to January 2018 were enrolled. According to the diagnostic criteria of kidney disease outcomes quality initiative (K/DOQI), patients were divided into CKD 1-5 stages. In all patients, Tc-GFR was measured by clearance rate of 99m Tc-diethylene triaminepen taacetic acid (99m Tc-DTPA) within 1 month, and SCr was determined. The eGFR was calculated by Cockcrofi-Gault (C-G) formula bases on SCr, Chinese modified modification of diet in renal disease (cMDRD) equation, Chronic Kidney Disease Epidemiology Collaboration Group (CKD-EPI) equation, measure and calculation (MC) equation and FAS equation, respectively, and body surface area (BSA) was used for correction of eGFR. The differences of Tc-GFR and eGFR calculated by different formulas were compared, and the correlations between eGFR calculated by different formulas and Tc-GFR were analyzed by Spearman correlation analysis. Tc-GFR was used as a reference standard to evaluate the bias, precision and accuracy of eGFR formulas, and the receiver operating characteristics (ROC) curve of each eGFR formula was plotted to evaluate its diagnostic efficacy for renal insufficiency in patients with CKD.

RESULTS: According to the inclusion and exclusion criteria, 382 patients with CKD were enrolled in the final analysis. There were 31, 69, 92, 75 and 115 patients with CKD 1-5 stages, respectively. In all patients, the differences between Tc-GFR and eGFR calculated by different formulas were statistically significant, and eGFR was positively correlated with Tc-GFR. The best correlation coefficient was between eGFR of cMDRD formula and Tc-GFR (r = 0.883, P = 0.000), and the lowest was of MC formula (r = 0.848, P = 0.000). The best correlation between the eGFR calculated by FAS formula and the Tc-GFR was in CKD 2 stage (r = 0.538, P = 0.000), and the lowest correlation was found in CKD 5 stage (r = 0.229, P = 0.014). Compared with Tc-GFR (the reference equation), the FAS formula showed the smallest bias [the difference between Tc-GFR and eGFR = 8.64, 95% confidence interval (95%CI) = 7.04-10.19], and the best accuracy [the percentage of eGFR falling into the range of Tc-GFR ±30% (P30) = 42.67%, 95%CI = 37.69-47.65]; CKD-EPI equation showed the best precision (QR of the difference between Tc-GFR and eGFR = 17.43, 95%CI = 15.33-21.28). ROC curve analysis showed that the area under the curve (AUC) of cMDRD formula was the largest (0.944), and the specificity was the highest (87.23%); the sensitivity of CKD-EPI formula was the highest (94.00%); and the AUC of MC formula was the smallest (0.918). The AUC of FAS formula was 0.940 (95%CI = 0.917-0.964, P = 0.000), it was higher than that of MC formula, but there was no significant difference between FAS formula and other formulas. When the cut-off value of eGFR calculated by FAS formula was 32.62 mL×min-1 ×1.73 m-2 , the sensitivity, specificity, positive predictive value, negative predictive value was 93.00%, 81.56%, 83.64%, 11.54%, respectively.

CONCLUSIONS: Compared with C-G formula, cMDRD formula, CKD-EPI formula and MC formula, FAS formula showed smaller bias and higher accuracy, and had higher specificity and sensitivity in the diagnosis of renal insufficiency in patients with CKD, which could be applied to the determination of GFR in early stage of CKD.

Clinical Trial Registration: Chinese Clinical Trial Registry, ChiCTR1800017727.