The study protocol was approved by Hebei Medical University ethical committee (NO. 2017-027-1), and the written informed consent was obtained from each participant.

Those subjects with the following criteria were identified and enrolled in the study: (1) Chinese patients; (2) at least 18 years of age; (3) the participants met the diagnostic standard for CKD.22 Patients with acute kidney function deterioration, edema, cardiac insufficiency, pleural or abdomen effusion, disabled limb, and treated with cimetidine or trimethoprim or replacement therapy were excluded.17,23–24 The patients were categorized into 2 subgroups based on the GFR calculated by reference method: lower-GFR subgroup (mGFR <60ml/min/1.73m2) and higher-GFR subgroup.

(mGFR>60ml/min/1.73m2).

The 99mTc-diethylene triamine pentaacetic acid (99mTc-DTPA) dual plasma sample clearance method was employed as the reference method. The main procedure was collecting the blood sample 2 and 4h after the 99mTc-DTPA injection. And then we measured the radioactivity of the blood sample immediately. The mGFR was calculated from a single exponential formula. The detailed procedure was described previously.17,23

The blood samples were collected between 6:30 and 8:00a.m. and then the blood samples were transferred to the laboratory within 2h. The blood samples were analyzed after they were centrifuged at 3000rpm for 15min. All the staffs in the laboratory underwent the same training and operated the biochemical analyzer according to the standard operation procedure. The serum creatinine (Scr) was automatically measured by the enzymatic IDMS-traceable method using a biochemical analyzer (VITROS 5.1, Johnson Company, USA; reagents from the same company).

The FAS equation was as follows8:

Q-values are the mean or median Scr value for age-/sex-specific healthy populations.10 The units of Scr and age were mg/dL and year respectively.

The Asian modified CKD-EPI equation is as follows10:

Female with the concentration of serum creatinine≤0.7mg/dl,

Female with the concentration of serum creatinine≥0.7mg/dl,

Male with the concentration of serum creatinine≤0.9mg/dl,

Male with the concentration of serum creatinine≥0.9mg/dl,

The unit of Scr and age was mg/dL and year respectively.

All Statistical analyses were performed using SPSS (version 17.0, SPSS. Chicago IL, USA) and Medcalc (version 4.3, Medcalc software, Mariekerke, Belgium). The normal distribution data was expressed by means of mean and standard deviation. The non-normal distribution data was expressed as median and interquartile range (M(QR)). The parameters of bias, precision, and accuracy were employed to compare the performance of the eGFR-FAS and eGFR-CKD-EPI. Bias was defined as eGFR minus mGFR. The standard deviation of the bias was considered as precision. The percentage of eGFR within 30% deviation of mGFR (P30) was employed as accuracy. The association between mGFR and eGFR was analyzed by means of the correlation analysis using the coefficient of Pearson. Wilcoxon Z test, F test were used to compare the bias, precision, accuracy and correlation coefficient. The receiver operating characteristic (ROC) curve illustrated the diagnostic ability for detecting the mGFR<60ml/min/1.73m2 of the two equations and the Z-test was employed to compare the area under curve (AUC).

A total of 160 patients (mean age 56±15 years) were enrolled in this study, including 76 males and 84 females. The values of serum creatinine, and mGFR were 1.96±1.65mg/dL, and 56.4±35.1ml/min/1.73m2, respectively. The eGFR-FAS and the eGFR-CKD-EPI were 55.2±30.4 and 57.7±34.3ml/min/1.73m2, respectively. The range that the values reached with the eGFR equations were less than those shown with mGFR. The mGFR reached 176.3ml/min/1.73m2 while eGFR-FAS only was 141.5ml/min/1.73m2. With eGFR-CKD-EPI it reached 138.6ml/min/1.73m2. A wide variety of clinical diagnoses led to chronic kidney disease were included in the study including chronic glomerulonephritis in 52 cases, diabetic nephropathy in 36 cases, chronic pyelonephritis in 30 cases, hypertensive nephropathy in 21 cases and other causes or unknown causes in the remaining 23 cases. The basic characteristics of the patients are shown in Table 1.

The scatter plots showed both eGFR-FAS and eGFR-CKD-EPI were highly associated with mGFR, and the correlation coefficients were 0.89 (P<0.001) and 0.89 (P<0.001). The linear regression was therefore constructed and the regression equations of eGFR-FAS and mGFR was eGFR-FAS=11.56+0.78mGFR (P<0.001). The regression equations of eGFR-CKD-EPI and mGFR was eGFR-CKD-EPI=9.32+0.87mGFR (P<0.001) (Figs. 1 and 2). Indeed, there was a close relationship between eGFR-FAS and Scr both in men and women (Fig. 3). Bland–Altman plots for the agreement of eGFR-FAS and eGFR-CKD-EPI against mGFR are depicted in Fig. 4. The limit of agreement for the eGFR-FAS and eGFR-CKD-EPI was 61.9 and 63.3ml/min/1.73m2, respectively. There was a tendency of both the equations to make greater errors when the values of mGFR were higher.

Fig. 1.

The scatter plot and linear regression of mGFR versus eGFR-FAS in 160 subjects. The line represents the regression line. Abbreviations: mGFR-the GFR measured by 99mTc-DTPA dual plasma clearance rate method; eGFR-FAS-GFR estimated by the FAS equation.

(0.08MB).

Fig. 2.

The scatter plot and linear regression of mGFR versus eGFR-CKD-EPI in 160 subjects. The line represents the regression line. Abbreviations: mGFR-GFR measured by the 99mTc-DTPA dual plasma clearance rate method; eGFR-CKD-EPI-GFR estimated by the Asian modified CKD-EPI equation.

(0.09MB).

Fig. 3.

The relationship between eGFR-FAS and Scr. The circle represents the female and the triangle represents the male. Abbreviations: eGFR-FAS-GFR estimated by the FAS equation; Scr-serum creatinine.

(0.05MB).

Fig. 4.

The Bland–Altman plots of difference versus mGFR in patients with CKD about the two equations. The mGFR are given on the X axis, and the difference of eGFR and mGFR is given on the Y axis. The solid long lines show mean difference and the dotted long lines represent the 95% limits of agreement. Abbreviations: eGFR-FAS-GFR: estimated by the FAS equation; eGFR-CKD-EPI-GFR: estimated by the Asian modified CKD-EPI equation.

(0.12MB).

In the whole cohort, the bias of eGFR-FAS and eGFR-CKD-EPI was −1.20ml/min/1.73m2 and 1.30ml/min/1.73m2 respectively. There was significant difference in the performance of bias (Z=−10.97, P<0.001). The precision of eGFR-FAS was not significantly lower than that of eGFR-CKD-EPI (15.9ml/min/1.73m2 vs 16.1ml/min/1.73m2; F=1.025, P=0.875). The eGFR-FAS and the eGFR-CKD-EPI also revealed the similar performance of accuracy (75.0% vs 76.3%, P=0.424). Regarding the correlation coefficient, the FAS equation did not have a superior performance compared with the Asian modified CKD-EPI equation (0.89 vs 0.89, Z=−0.087, P=0.930).

The AUCs of were 0.94 (0.90–0.97) for FAS equation, 0.94 (0.90–0.97) for Asian modified CKD-EPI equation. There were no significant differences in the abilities (AUCs) of the equations to detect the mGFR less than 60ml/min/1.73m2 (Z=0.221, P=0.825) (Fig. 5). When the sensitivity and specificity were most satisfactory, the cut-off value of eGFR-FAS was 60.8ml/min/1.73m2 (sensitivity: 92.3%; specificity: 82.6%) and the cut-off value of eGFR-CKD-EPI was 60.3ml/min/1.73m2 (sensitivity: 84.6%; specificity: 92.8%).

Fig. 5.

The Receiver Operating Characteristic curve of the two equations for the diagnostic ability for detecting the mGFR<60ml/min/1.73m2. The green curve represented the diagnostic ability of FAS equation and the red curve represented the diagnostic ability of Asian modified CKD-EPI equation.

(0.11MB).

The performance of bias, precision and accuracy in the mGFR<60ml/min/1.73m2 subgroup and mGFR>60ml/min/1.73m2 subgroup were consistent with that in the whole cohort. (The comparison of the two equations is shown in Table 2.)