Novel Diagnostic and Prognostic Urinary Biomarker Model for Lupus Nephritis and Renal ANCA-Associated Vasculitis

Objectives The gold standard for the diagnosis of lupus nephritis (LN) and renal ANCA-associated vasculitis (rAAV) is renal biopsy, an invasive procedure attended by the risk of bleeding and other complications. While current biomarkers such as proteinuria and hematuria are useful, they are not always reflective of disease activity. This study evaluates the diagnostic and predictive performance of urinary soluble CD163 (usCD163), urinary complement activation products (uCAPs), and urinary SIGLEC-1 as non-invasive biomarkers of renal inflammation in LN and rAAV. Methods Urine samples and kidney biopsy data were from the Biobank for Molecular Classification of Kidney Disease from patients with LN (n=14), rAAV (n=10), and healthy controls (n=10). Samples ranged from 14 days pre-biopsy to 238 days post-biopsy. Biomarker levels were quantified using ELISA (usCD163, sC5b9, SIGLEC-1) and Meso Scale Discovery (C3a, C5a) assays and normalized to urine creatinine levels. Mean concentrations were compared between groups, logistic regression generated a combined model, and receiver operating characteristic (ROC) curves assessed performance in predicting complete renal remission (CRR; UPCR<300 mg/g). Results UsCD163 and urine-protein creatinine ratio (UPCR) levels were significantly elevated in patients with LN (17.92 ± 14.09 and 241.21 ± 74.41 mg/mmol, respectively) and rAAV (4.27 ± 2.03 and 90.87 ± 30.40 mg/mmol, respectively) compared to healthy controls (0.13 ± 0.01 and 7.53 ± 1.67, respectively), whereas uCAPs and SIGLEC-1 were not. There was a strong positive correlation between UPCR and usCD163 (ρ=0.886, p<0.001), and moderate correlations with C3a (ρ=0.642, p<0.001). C5a (ρ=0.588, p<0.001), and sC5b9 (ρ=0.566, p=<0.001). SIGLEC-1 concentrations showed no significant correlation with UPCR (ρ=0.181, p=0.31). A combined model including usCD163 and uCAPs achieved strong predictive performance for distinguishing patients who achieved complete renal remission from those who did not (AUC=0.91; Figure 1). Among individual biomarkers, usCD163 had the highest discriminative performance (AUC=0.96), while sC5b-9 had the lowest (AUC=0.74; Figure 1). Figure 1. Receiver operating characteristic (ROC) curve for the prediction of complete renal remission (CRR) . ROC analysis was performed for each biomarker- usCD163 (AUC = 0.96), C3a (AUC = 0.82), C5a (AUC = 0.80), and sC5b9 (AUC = 0.74)- as well as a combined model (AUC = 0.91). The combined model, including usCD163 and uCAPS, demonstrated strong predictive performance of distinguishing patients with CRR from non-CRR, with sCD163 having the highest individual discriminative accuracy. Conclusion Urinary biomarkers reflecting distinct aspects of renal immune activation have potential to non-invasively monitor disease activity in LN and rAAV. Our findings support usCD163 as an indicator of renal inflammation, correlating closely to disease activity and remission status. A combined model with uCAPs and UPCR improved diagnostic performance further, capturing key immunologic differences in disease pathogenesis. Further validation in a larger, longitudinal cohort is underway.