Comparisons Between The Glomerular Filtration Rate Estimating Equations Reported By CKD-EPI In 2012 And Other Two Equations

Objective:Chronic Kidney Disease Epidemiology Collaboration reported two new equations in 2012:one based on cystatin C alone(CKD-EPIcys), the other using both creatinine and cystatin C(CKD-EPIcr-cys). The study compared the adaptability of the two new equations with the other two classic equations(simplified MDRD and CKD-EPIcr) in CKD patients and healthy kidney donors. To exploring whether the two new equations are applicable to the people in our area.Methods:The study selected 304 patients from January 2013 to June 2015 in nephrology and kidney transplantation center of The First Hospital of Jilin University, we divided the patients into five groups by the level of r GFR according to K/DOQI guide:CKD1(≥90ml/min/1.73m2),CKD2(60-89ml/min/1.73m2),CKD3(30-59ml/min/1.73m2),CKD4(15-29ml/min/1.73m2),CKD5(<15ml/min/1.73m2). 99 m Tc-DTPA clearance by renal dynamic imaging method was used as reference standard. Bias, absolute bias,precision, accuracy(P20,P30) were used to compare the performance of the four equations in different periods, correct classification proportion and error estimated percentage of CKD using e GFRs were also compared. Meanwhile, all the patients were divided into Elderly(≥60 years) and Non-Elderly(<60 years), the adaptability of the two groups would be compared.Results:1. In the overall sample, the four equations all existed different levels of underestimation, the correlation and precision of CKD-EPIcys and CKD-EPIcr-cys are superior to the other two equations, accuracy(P20,P30) is higher than the other two equations, P20 of CKD-EPIcys is higher than MDRD, but no significant difference with CKD-EPIcr(p>0.05), P30 is higher than CKD-EPIcr, but no significant difference with MDRD(p>0.05).2. Comparisons bettween the four equations under the different stages2.1 CKD1 The absolute bias of the two new equations are both smaller than MDRD, bias and absolute bias of CKD-EPIcr-cys are both smaller than CKD-EPIcr and CKD-EPIcys, although the accuracy(P20,P30) of the four equations do not have significant difference(p>0.05), P20 of the two new equations are both higher than the other two classical equation, accuracy(P20,P30) exceeded 80%.2.2 CKD2 The absolute bias of CKD-EPIcys is smaller than MDRD, bias andabsolute bias of CKD-EPIcr-cys are smaller than MDRD and CKD-EPIcys,although the accuracy(P20,P30) of the four equations do not have significantdifference(p>0.05), P30 exceeded 70%.2.3 CKD3 The bias and absolute bias of the two new equations are bothsmaller than MDRD and CKD-EPIcr, P20 is higher than CKD-EPIcr,accuracy(P20,P30) of CKD-EPIcr-cys are higher than MDRD.2.4 CKD4-5 The bias and absolute bias of the two new equations are bothsmaller than the other two equations, accuracy(P20,P30) of CKD-EPIcr-cys areboth higher than MDRD and CKD-EPIcr.3. Comparisons of the percentages that were classified in the correct CKD stages by the e GFRs from the four equationsChi-square test compared the consistency of CKD stages between the four equations, CKD2, CKD4 and CKD5, the difference was statistically significant(p<0.05), in CKD2, correct classification proportion of MDRD is lower than the other equations, in CKD4, correct classification proportion of CKD-EPIcys is higher than MDRD and CKD-EPIcr, in CKD5, correct classification proportion of CKD-EPIcr-cys is higher than the other equations, overall, correct classification proportion of the two new equations are higher than the other two equations.4. Comparisons of the agreement between classification of CKD stages and the percentages that were classified in the wrong CKD stages by the e GFRs from the four equationsThe result of Kappa test showed that the agreement between classification of CKD stages by e GFRs calculated by the four equations and that by r GFR is not good,Kappa values are beween 0.41 and 0.60.Chi-square test compared the misclassified percentages of CKD stages, in CKD2,MDRD underestimates the r GFR, that resulted in 60.46% of the patients were misclassified into CKD3 and CKD4, the levels of misestimation of the other threeequations decreased obviously, the difference was statistically significant(p<0.05).Likewise, in CKD3, 51.35% of the patients were underestimated when e GFR calculated by MDRD, compared with MDRD, CKD-EPIcr-cys was improved, the difference was statistically significant(p<0.05). In CKD4, 75% of the patients were underestimated when e GFR calculated by MDRD, there was significant difference with CKD-EPIcr-cys(p<0.05).In CKD1, 14.49% of the patients were underestimated when e GFR calculated by CKD-EPIcys, the levels of underestimation has been improved, there was signicant difference with CKD-EPIcys(p<0.05).5. Comparisons of indicators between Elderly group and Non-Elderly groupIn r GFR<60 ml/min/1.73m2 and r GFR≥60 ml/min/1.73m2, Elderly absolute bias is larger than Non-Elderly, but no significant difference(p>0.05), precision is not improved, either. In r GFR<60 ml/min/1.73m2, Elderly P30 has no significant difference with Non-Elderly(p>0.05); In r GFR≥60 ml/min/1.73m2, Non-Elderly P30 is significantly higher than Elderly(p<0.05), exceeded 80%. In terms of absolute bias and precision, no matter Elderly of Non-Elderly, the two new equations are superior to MDRD.Conclusions:1. In terms of correlation, bias, absolute bias, precision and accuracy, the two new equations are superior to MDRD and CKD-EPIcr, are suitable to the patients in all CKD stages.2. CKD-EPIcr-cys is better than CKD-EPIcys, especially r GFR≥60ml/min/1.73m2,bias and absolute bias of CKD-EPIcr-cys are better than CKD-EPIcys, in CKD1, CKD-EPIcr-cys the levels of underestimating r GFR is improved, compared with CKD-EPIcys.3. The adaptability of the four new equations are all not perfect for the Elderly,in r GFR≥60 ml/min/1.73m2, Non-Elderly P30 is significantly higher than Elderly,the two new equations are suitable to CKD patients and healthy kidney donors.