|
ObjectivesThere is no data about glomerular filtration rate, measured through plasma clearance of the exogenous markers, in the Chinese children aged from1-to18-year-old. The objectives of this study include:(1) to measure GFR and validate the nine GFR estimating equations in the Chinese children, aiming at selecting the equations with best performance in GFR estimation and CKD staging;(2) to determine the threshold values of serum creatinine and serum cystatin C for diagnosing impaired renal function in different age groups;(3) to validate the performance of the five GFR estimating equations that are used in adults in the subgroup of children aged14or more;(4) to compare the nine GFR estimating equations in the subgroup of children with the diagnosis of systemic lupus erythematosus.MethodsSingle-compartment plasma clearance of99mTc-DTPA was measured using the radioactive counts of the plasma sampled at two different time spots after the peripheral bolus injection of measured dose of99mTc-DTPA and transformed into two-compartment plasma clearance, which was used as measured GFR (mGFR) for subsequent analysis, using the Brochner-Mortenson equation. mGFR was normalized by body surface area, which was calculated using the Haycock equation, and nGFR was used for correlation, agreement and accuracy analysis. Univariate linear regression was used to determine the capacity of markers, including those related to body size and those related to renal function, in explaining the inter-individual variance of mGFR and nGFR. eGFR was calculated using nine different equations, of which three are based on sCr, three based on sCysC and three based on the combination of sCr and sCysC. Pearson correlation coefficient and univariate linear regression were used to compare the correlation between eGFR from different equations and nGFR. Bias,95%limits of agreement, Bland-Altman analysis, intraclass correlation coefficient and concordance correlation coefficient were used to compare the agreement between eGFR from different equations and nGFR. The ratio of eGFR within nGFR±10%and nGFR±30%, the ratio of correct CKD staging by eGFR and Kappa analysis were used to compare the accuracy of GFR estimation of different equations. Receiver operating curve analysis was performed and the diagnostic threshold of sCr and sCysC were determined with equal weight in sensitivity and specificity. Five adult GFR estimating equations were used to calculated eGFR in the children aged over14and compared with the pediatric equations, using the similar analyzing process as that used in the whole group. Children with diagnosis of systemic lupus erythematosus was selected and the nine pediatric GFR estimating equations were compared in this subgroup, using the similar analyzing process as that used in the whole group.Results1. In the whole group (N=85), compared with the other six equations, Equation-Schwartz1976, E-Filler and E-CKiD produced the eGFR with better correlation with nGFR (γpearson was0.60,0.73and0.75, respectively), stronger explanation capacity of inter-individual variance in nGFR (R2for univariate linear regression model was0.612,0.716and0.730, and RMSE was0.116,0.094and0.092, respectively), smaller bias (bias was0.2,-4.2and3.5[unit:ml/min x1.73m2], respectively), narrower95%LOA ([-60.0,61.0]、[-50.1,41.7] and [-37.1,44.1] respectively), better performance in Bland-Altman analysis (slope of linear regression line was0.06,-0.16,0.10, respectively), higher ICC (0.594,0.724and0.741, respectively), higher CCC (0.592,0.722and0.735, respectively), higher ratio of eGFR within nGFR±10%(30.6%,33.8%and37.7%, respectively) and eGFR within nGFR±30%(81.2%,76.6%and77.9%, respectively), higher ratio of correct CKD staging (65.9%,68.8%and75.3%, respectively) and better agreement in CKD staging between eGFR and nGFR (Kappa value was0.61,0.61and0.65, respectively).2. In each of the three age groups (1-7y/o,7.1-14y/o and14.2-18y/o), the best diagnostic values of sCr, determined by ROC analysis,, were0.55,0.75and1.07mg/dL for nGFR≤60ml/min x1.73m2, respectively;0.42,0.55and0.66mg/dL for nGFR≤90ml/min x1.73m2, respectively. The best diagnostic values of sCysC, determined by ROC analysis, were1.91,1.64and1.64mg/dL for nGFR≤60ml/min x1.73m2, respectively;0.92,1.05and1.07mg/dL for nGFR≤90ml/min x1.73m2, respectively.3. In the children aged over14(N=29), five adult equations (E-CG, E-simplified MDRD, E-modified MDRD, E-CKD EPI-Creatinine, E-CKD EPI Combination) were used to calculate eGFR. Compared with the nine pediatric equations, the adult equations produced eGFRs with weaker correlation, poorer agreement and lower ratio of eGFR within nGFR±10%and nGFR±30%.4. In the children with the clinical or pathological diagnosis of systemic lupus erythematosus (N=24), E-Schwartz1976and E-CKiD, compared with the other seven pediatric equations, produced eGFR with better correlation with nGFR (γPearson was0.58and0.57, respectively), smaller bias (1.8and4.12ml/min×1.73m2, respectively), narrower95%LOA ([-29.0,31.8] and [-19.5,39.7], respectively), higher ICC (0.55and0.55, respectively), higher CCC (0.54and0.53, respectively), higher ratio of eGFR within nGFR±10%(33.3%and42.9%, respectively) and eGFR within nGFR±30%(83.3%and81.0%, respectively), better agreement in CKD staging between eGFR and nGFR (Kappa value was0.45and0.47, respectively) and higher ration of correct CKD staging (58.3%and71.4%, respectively).Conclusions1. The selected equations used to estimate GFR in the whole study group were E-Schwartz1976, E-Filler and E-CKiD;2. Significant differences existed in the best diagnostic value of sCr for impaired renal function between different ages groups. The impact of age on the diagnostic value of sCysC for renal function impairment was non-significant.3. The five adult equations can not be applied to the children aged over14to estimate GFR;4. E-Schwartz1976and E-CKiD are the best equations to estimate GFR in the children with the diagnosis of systemic lupus erythematosus;
|