Predict Children With Primary Nephrotic Syndrome Of Urinary Biomarkers Of Hormone Curative Effect Of Study

Objective To detect the urinary AAT, PA and ZAG concentrationsin children with primary nephrotic syndrome （PNS） before glucocorticoid treatmentand to verify whether these parameters could predict the response of glucocorticoidtherapy. Methods43children diagnosed PNS initially were chosen as subjects,namely SSNS and SRNS depending on reaction to glucocorticoid therapy four weekslater, and15healthy children serving as normal control. The mid stream of the firstmorning urine samples were collected from children before taking glucocorticoid.ELISA kit was used to quantify the AAT, PA and ZAG concentrations of the urinesamples which were revised by urine creatinine further. Data analysis was performedusing the SPSS17.0. Results1. AAT was absent in normal urine samples. Thelevel of UAAT/Cr in children with SRNS was higher than that in children withSSNS[0.049（0.028⁰.073） vs0.028（0.022⁰.036）, Z=2.4, P=0.02].2. The detectionvalue of PA was58.0（41.0⁷4.1）ng/ml and the value of UPA/Cr was5.38（3.16⁹.37）×10-5in normal healthy children. The level of UPA/Cr in children with SRNS washigher than that in children with SSNS[3.87（2.04⁷.44）×10-3vs1.38（0.24².65）×10-3,Z=-3.7, P=0.000].3. ZAG was tested negative in normal urine. The level ofUZAG/Cr has no statistic difference between children with SRNS and SSNS（t=1.4,P=0.16）.4. There were no statistically significant differences of the levels ofUAAT/Cr, UPA/Cr and UZAG/Cr among the different pathological types of thechildren who had renal biopsy. Conclusions The level of UAAT/Cr and UPA/Crwere higher in the SRNS that can be as candidate biomarkers to predict the response of glucocorticoid-based therapy. Objective To construct a model for predicting response toglucocorticoid therapy with biomarkers which have statistic differences betweenchildren with SSNS and SRNS. Methods The laboratory parameters whichincluded RBC, HB, PLT, WBC, N%, L%, E%, TB, ALT, AST, ALP, CysC, BUN,SCr, UA, TP, ALB, GLB, TG, CHOL, HDL, IgG, IgA, IgM, C3, TS%, TH%, NK%and B%of the blood and RBC, WBC, MA, IgG, A1M, UPro/UCr, UCr and24hUPro/kg of the urine were collected from43children with PNS before treatmentof glucocorticoid. Parameters that were found to be statistically significant andurinary protein biomarkers in partⅠwere selected for further logistic regressioninvestigation. The prediction model was constructed via logistic regression analysisand ROC was performed to predict the efficacy of the model using the SPSS17.0.Results1. Among the laboratory parameters of the two subgroups before takingglucocortiod, the levels of platelet, blood white cell counts, serum globulin, urinewhite cell counts, urine red cell counts, urine IgG and urine a1-microglobulin havesignificant statistic differences（P<0.05）.2. Four parameters that included serumglobulin（OR=1.9, P=0.03）, UA1M（OR=1.0, P=0.03） UAAT/Cr (OR=1.2×10¹⁸,P=0.16) and UPA/Cr （OR=1.9, P=0.03） further entered the logistic regression model(The logistic equation: P=e^{-23.147+1.181GLB+1.785UA1M+0.764UAAT/Cr+1.478UPA/Cr}/1+e^{-23.147+1.181GLB+1.785UA1M+0.764UAAT/Cr+1.478UPA/Cr}) to predict the SRNS independently.3. The ROC curves based on eachparameter which was in the logistic regression model were performed and the areasunder the curves（AUC） were0.69⁰.83. The sensitivities and specificities were63%⁷4%and63%⁹2%respectively （Youden’s indexes0.35⁰.55）. The AUC thatbased on the logistic regression model which included serum globulin, UA1M,UAAT/Cr and UPA/Cr was improved to0.97, and the sensitivity and specificity of themodel prediction were95%and96%respectively （Youden’s index0.91）.Conclusions The AUC that based on the logistic regression model which includedserum globulin, UA1M, UAAT/Cr and UPA/Cr was improved to0.97. When the Pvalue of the logistic equation was greater than or equal to0.56, the sensitivity andspecificity of the model prediction were95%and96%respectively （Youden’s index0.91）. It had a better prediction efficacy based on the model which included serumglobulin, UA1M, UAAT/Cr and UPA/Cr compared to one single biomarker.