URINARY ADHESION MOLECULES PREDICT HISTOLOGIC SEVERITY AND RENAL OUTCOMES IN ANCA-ASSOCIATED GLOMERULONEPHRITIS

Kidney biopsy provides valuable histological information but is invasive and carries a bleeding risk in patients with ANCA-associated glomerulonephritis (ANCA-GN). Therefore, noninvasive urinary biomarkers that reflect renal pathology and prognosis are desirable. Intercellular adhesion molecule 1 (ICAM-1) and vascular cell adhesion molecule 1 (VCAM-1) belong to the immunoglobulin superfamily and act as cell adhesion molecules. These molecules are normally expressed at low levels, commonly on vascular endothelial cells, and are upregulated during inflammation, which facilitates leukocyte recruitment and contributes to vasculitic injury. Recent studies have identified a proximal tubule subset with high VCAM-1 expression during recovery after acute kidney injury and have linked tubular VCAM-1 to tubular injury and the progression from acute kidney injury to chronic kidney disease. To our knowledge, these changes have not been reported in ANCA-GN. However, similar mechanisms may operate in ANCA-GN, and urinary levels of these adhesion molecules may reflect the severity of kidney injury and prognosis. The aim of this study is to assess the clinical significance of urinary ICAM-1 and VCAM-1 in relation to histopathology and the longitudinal change in kidney function (eGFR) after treatment initiation in ANCA-GN.

Patients with ANCA-GN who underwent kidney biopsies in an institutional cohort (F-KDR, n = 76) and a nationwide cohort (RemIT-JAV-RPGN, n = 44) were included. Urinary ICAM-1 and VCAM-1 levels at biopsy were quantified by ELISA, and relationships with histological findings were evaluated using Wilcoxon rank-sum tests, receiver operating characteristic curve analyses and logistic regression. In the F-KDR cohort, we first conducted univariable analyses of UVIR. A multivariable ridge-penalized model was then fit to assess its independent association and incremental value beyond the baseline model. External validation was performed in the RemIT-JAV-RPGN cohort. eGFR trajectories were modeled using a linear mixed-effects model in the F-KDR cohort, and a mixed model for repeated measures in the RemIT-JAV-RPGN cohort.

The urine VCAM-1/ICAM-1 ratio (UVIR) was strongly associated with interstitial fibrosis and tubular atrophy (IF/TA) severity, whereas neither ICAM-1 nor VCAM-1 alone showed a clear association with IF/TA severity. For UVIR, the AUCs (95% CIs) for discriminating IF/TA ≥25% and ≥50% were 0.87 (0.79–0.95) and 0.85 (0.77–0.94) in the F-KDR cohort, and 0.76 (0.62–0.90) and 0.77 (0.60–0.95) in the RemIT-JAV-RPGN cohort. In the F-KDR cohort, univariable logistic regression yielded odds ratios (ORs) for log(UVIR) of 9.01 (95% CI 3.54–31.95) for IF/TA ≥25% and 5.21 (2.52–13.54) for IF/TA ≥50%. In multivariable ridge-penalized logistic regression, log(UVIR) remained an independent predictor (Table). External validation in the RemIT-JAV-RPGN cohort yielded AUCs of 0.80 (0.67–0.93) for IF/TA ≥25% and 0.82 (0.68–0.96) for IF/TA ≥50%. Across both cohorts, eGFR trajectories modeled with UVIR were similar to those obtained when stratifying by IF/TA severity (Figure).

The UVIR is a reliable and promising biomarker that reflects IF/TA severity and is associated with eGFR trajectory in ANCA-GN.