Detection of urinary α1-antitrypsin for predicting response to glucocorticoid therapy in children with primary nephrotic syndrome / 中华肾脏病杂志

ABSTRACT

Objective To detect the α1-antitrypsin (AAT) concentration in urine samples of children with primary nephrotic syndrome (PNS) before initiation of glucocorticoid treatment,in order to verify whether it could predict the response to glucocorticoid-based therapy.Methods Forty-three children diagnosed as PNS initially were chosen as subjects,namely steroid-sensitive nephrotic syndrome (SSNS) and steroid-resistant nephrotic syndrome (SRNS) depending on reaction to glucocorticoid therapy four weeks later,and 15 healthy children serving as normal control.The mid stream of the first morning urine samples were collected from children before taking glucocorticoid.ELISA kit was used to quantify the urinary AAT concentration which was revised by urine creatinine further.The data of urine AAT/Cr were expressed as median with interquartile range.Data analysis was performed using the SPSS 17.0.Results AAT was absent in urine samples of normal healthy children,and there were no statistic differences of the AAT concentrations in urine between children with SSNS and SRNS [(30.4+4.5) mg/L vs (31.8+4.6) mg/L,t=-1.0,P=0.33].The level of urine AAT/Cr in children with SRNS was higher than that in children with SSNS [0.049(0.028-0.073) vs 0.028(0.022-0.036),Z=2.4,P=0.02].Among the laboratory parameters of the two subgroups before taking glucocortiod,the levels of platelet,blood white cell count,serum globulin,urine white cell count,urine red cell count,urine IgG and urine α1-microglobulin were significantly different (P＜0.05).Three parameters that included urine AAT/Cr (OR=6.81 × 1028,P=0.O05),serum globulin (OR=1.69,P=0.01) and urine α1-microglobulin (OR=1.05,P=0.009) further entered the logistic regression model to predict the SRNS independently.The ROC curve based on the level of the urine AAT/Cr was constructed,and the area under the curve (AUC) was 0.72.When the cutoff value of urine AAT/Cr was 0.035,the sensitivity and specificity of the urine AAT/Cr prediction were 68％ and 75％ respectively (Youden' s index 0.43).The AUC that based on the logistic regression model which included urine AAT/Cr,serum globulin and urine α1-mieroglobulin was improved to 0.94,and the sensitivity and specificity of the model prediction were 95％ and 83％ respectively (Youden' s index 0.78).There was no significant difference of the urine AAT/Cr level among the different pathological types of the children undergoing renal biopsy.Conclusions There are no statistic differences of the AAT concentrations in urine between children with SSNS and SRNS.The level of urine AAT/Cr is significantly higher in the SRNS than that in the SSNS which can be as a candidate biomarker to predict the response to glucocorticoid-based therapy.It has a better prediction efficacy based on the model which includes urine AAT/Cr,serum globulin and urine α1-microglobulin.