BACKGROUND: The new Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation was developed to address the systematic underestimation of glomerular filtration rate (GFR) by the Modification of Diet in Renal Disease (MDRD) Study equation in patients with relatively well-preserved kidney function. Performance of the new equation in the Chinese population is unknown. The goal of the present study was to compare performance of these two equations in Chinese patients with chronic kidney disease (CKD). METHODS: We enrolled 450 Chinese patients (239 women and 211 men) with CKD in the present study. The renal dynamic imaging method was used to measure the referenced standard GFR (rGFR) for comparison with estimations using the two equations. Their overall performance was assessed with the Bland-Altman method and receiver-operating characteristics (ROC) analysis. Performance of the two equations in lower and higher estimated GFR (eGFR) subgroups was further investigated. RESULTS: Both eGFRs correlated well with rGFR (r = 0.88, 0.81, P < 0.05). In overall performance, the CKD-EPI equation showed less bias, higher precision and improved accuracy, and was better for detecting CKD. In the higher-eGFR subgroup, the CKD-EPI equation corrected the underestimation of GFR by the abbreviated MDRD equation. CONCLUSIONS: The CKD-EPI equation outperformed the abbreviated MDRD equation not only in overall performance but also in the subgroups studied. For the present, the CKD-EPI equation appears to be the first-choice prediction equation for estimating GFR.
BACKGROUND: The new Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation was developed to address the systematic underestimation of glomerular filtration rate (GFR) by the Modification of Diet in Renal Disease (MDRD) Study equation in patients with relatively well-preserved kidney function. Performance of the new equation in the Chinese population is unknown. The goal of the present study was to compare performance of these two equations in Chinese patients with chronic kidney disease (CKD). METHODS: We enrolled 450 Chinese patients (239 women and 211 men) with CKD in the present study. The renal dynamic imaging method was used to measure the referenced standard GFR (rGFR) for comparison with estimations using the two equations. Their overall performance was assessed with the Bland-Altman method and receiver-operating characteristics (ROC) analysis. Performance of the two equations in lower and higher estimated GFR (eGFR) subgroups was further investigated. RESULTS: Both eGFRs correlated well with rGFR (r = 0.88, 0.81, P < 0.05). In overall performance, the CKD-EPI equation showed less bias, higher precision and improved accuracy, and was better for detecting CKD. In the higher-eGFR subgroup, the CKD-EPI equation corrected the underestimation of GFR by the abbreviated MDRD equation. CONCLUSIONS: The CKD-EPI equation outperformed the abbreviated MDRD equation not only in overall performance but also in the subgroups studied. For the present, the CKD-EPI equation appears to be the first-choice prediction equation for estimating GFR.