Diagnostic value of combined detection of urine and plasma kidney injury markers in children with Henoch-Sch?nlein purpura nephritis / 中华检验医学杂志

ABSTRACT

Objective:To investigate the clinical significance and the diagnostic value of detecting kidney injury biomarkers in urine and serum of children with Henoch-Sch?nlein purpura nephritis (HSPN).Methods:A total of 216 children with untreated HSPN, who were admitted in Beijing Children′s Hospital of Capital Medical University from January 2018 to December 2019, were recruited in this retrospective study. Two hundred and sixteen healthy children were selected as the healthy control group. We determined the levels of six biomarkers of kidney injury, including transferrin (TRF), immunoglobulin (IgG), microalbumin (mAlb), alpha-1 microglobulin (α1-MG), N-acetyl-β-D-glucosaminidase (NAG) in urine and cystatin C (CysC) in serum. The data from the two groups were analyzed, the diagnostic value of each biomarker was evaluated and a logistic regression model for the diagnosis of HSPN was established. In addition, 60 children with HSPN, who were admitted to our hospital from November 2021 to February 2022 and 60 healthy children, who underwent healthy check up in the same period were included to validate the diagnostic performance of the established logistic model. Receiver operating characteristic (ROC) curve was used to analyze the diagnostic value of each biomarker.Results:The urine levels of TRF, IgG, mAlb, α1-MG and NAG and the serum level of CysC were significantly higher in the HSPN group than those in healthy control group (all P<0.05). The area under the ROC curve (AUC) of TRF, IgG, mAlb, α1-MG, NAG and the serum levels of CysC was 0.749, 0.719, 0.810, 0.648, 0.828 and 0.790 (all P<0.05). Logistics regression analysis showed that IgG, mAlb and TRF were the three diagnostic determinants of HSPN ( OR=1.083, 1.105, 1.704,all P<0.001), and the AUC was 0.916 of the established logistic model based on these three biomarkers. The sensitivity was 87.4% and the specificity reached 96.2%. The logistic model was validated by independent cohorts, and the AUC was 0.973, the sensitivity was 95.0% and the specificity was 98.3%. Conclusions:The levels of urine TRF, IgG, mAlb, α1-MG, NAG and serum CysC were higher in children with HSPN. The established logistic regression model based on three biomarkers including IgG, mAlb and TRF in this study has satisfactory clinical value in diagnosing HSPN in children.