Background/Purpose: Early detection of renal involvement in diabetic kidney disease (DKD) and lupus nephritis (LN) is critical but remains challenging. Conventional renal biomarkers (serum creatinine, eGFR, albuminuria) are often insensitive to subclinical damage and rise late in disease. We identified four kidney-specific plasma proteins (uromodulin [UMOD], myo‑inositol oxygenase [MIOX], heparanase [HPSE], and AREL1) linked to distinct renal processes: tubular transport, metabolic stress, extracellular matrix remodeling, and immune signaling. We hypothesized that an index combining these mechanistically distinct markers could detect early renal pathology in DKD and LN, outperforming conventional measures.

Methods: Plasma UMOD, MIOX, HPSE, and AREL1 were measured by immunoassay in three patient cohorts: 1) biopsy-confirmed active LN (n=31), compared to SLE without active proteinuria (n=31) and matched apparently healthy volunteer controls (AHV, n=31), ; 2) early-stage type 1 DKD obtained from the NIH-sponsored PERL study (n=49, mean eGFR=66 mL/min/1.73m²); and 3) late-stage type 2 DKD obtained from the FIND study (n=50, mean eGFR≈21 mL/min/1.73m²). The latter two cohorts were compared to matched AHVs (total AHVs=99). Logistic regression models using these four biomarkers were trained to distinguish DKD and LN from matched AHVs, generating DKD and LN indices (predicted probabilities). ROC analyses compared biomarker indices against eGFR, including comparisons of early DKD stratified by iohexol-determined GFR (iGFR), cystatin C, and albuminuria. AHV versus LN were stratified by urinary protein-to-creatinine ratio (UPCR). Data from PERL and FIND were provided by NIDDK CR, a program of the National Institute of Diabetes and Digestive and Kidney Diseases.

Results: The LN Index strongly differentiated LN from AHVs (AUC=1.00) and from non-proteinuric SLE (AUC=0.99), significantly outperforming eGFR (AUC=0.27 for LN vs AHV, AUC=0.48 for LN vs non-proteinuric SLE). The DKD Index robustly discriminated both early DKD (PERL, AUC=0.92) and late DKD (FIND, AUC=0.97) from AHVs, surpassing eGFR performance (AUC=0.81 early DKD, AUC=0.94 late DKD). Subgroup analyses demonstrated superior sensitivity of biomarker indices compared to eGFR for detecting early kidney damage defined by iGFR, cystatin C, or albuminuria in DKD. The LN index also outperformed UPCR in LN, while eGFR values mostly remained in the normal range.

Conclusion: A plasma biomarker index combining UMOD, MIOX, HPSE, and AREL1 detects DKD and LN with high sensitivity and specificity, identifying organ damage earlier and more clearly than standard measures. These kidney-specific proteins capture distinct pathophysiologic axes of renal damage, supporting their use as early, disease-specific indicators of nephropathy in diabetes and lupus. Such a multi-marker approach could improve timely diagnosis and monitoring of renal involvement beyond conventional tests.

Table 1. Demographic and Clinical Characteristics of Study Cohorts

Figure 1. Receiver Operating Characteristic (ROC) curve analysis of biomarker performances

Figure 2. Comparative analysis of eGFR and biomarker-derived indices (DKD index, LN index) across various kidney disease stratifications

« Back to ACR Convergence 2025

ACR Meeting Abstracts - https://acrabstracts.org/abstract/kidney-specific-four-protein-index-outperforms-conventional-measures-for-detecting-early-kidney-damage-in-lupus-nephritis-and-diabetic-kidney-disease/