A comparative study on the applicability of three CKD-EPI equations for estimation of glomerular filtration rate in Chinese patients with diabetic nephropathy

Objective To evaluate whether three chronic kidney disease epidemiology collaboration (CKD-EPI) equations (CKD-EPI2009Scr, CKD-EPI2012SCysC and CKD-EPI2012Scr-SCysC) are applicable in the prediction of glomerular filtration rate (GFR) in Chinese patients with diabetic nephropathy (DN). Methods One hundred and eight patients with DN who were hospitalized in the First Affiliated Hospital of Guangzhou Medical University with GFR being measured by dynamic renal imaging with 99mTc-DTPA from June 2012 to April 2014 were enrolled in this study. GFR measured by dynamic renal imaging with 99mTc-DTPA was used as the reference value (rGFR). GFR was estimated by the CKD-EPI2009Scr equation, the CKD-EPI2012SCysC equation, and the CKD-EPI2012Scr-SCysC equation (labeled as eGFR1, eGFR2, eGFR3). The correlation, 30% accuracy, staging consistency, deviation and diagnostic accuracy were compared among the three CKD-EPI equations. Results The rGFR in 108 DN patients was (61.78±26.51) ml·min-1·(1.73 m2)-1. The correlation between three eGFRs and rGFR was significant (all P 0.05), eGFR1 underestimated rGFR, eGFR2 overestimated rGFR (all P <0.01). The results of the Bland-Altman chart showed that consistencies between three eGFRs and rGFR were poor, the degree of deviation of eGFR3 was the smallest. The area under the ROC curve (AUC) of three eGFRs was 0.878, 0.883 and 0.915. The AUC, sensitivity, specificity, overall compliance rate and Youden index of eGFR3 were the highest. Conclusions The eGFRs predicted by the three CKD-EPI equations showed good relevance, accuracy and diagnostic accuracy with the rGFR, but poor in consistencies. Comparatively, CKD-EPI2012Scr-SCysC may be better than others, but its consistency limits exceeds the acceptable limits. Therefore, the applicability of using the three CKD-EPI equations to predict the GFR in Chinese DN patients requires a larger sample and multiple verifications as well as further improvement. Key words: Diabetic nephropathies; Kidney failure, chronic; Glomerular filtration rate; CKD-EPI equations