Background/Purpose Lupus Nephritis (LN) occurs in up to 80% of childhood-onset Systemic Lupus Erythematosus (cSLE) and it has a worse prognosis than adults. The current gold standard for diagnosing LN and assessing its activity is a kidney biopsy interpreted using the International Societies for Nephrology & Renal Pathology (ISN/RPS) classification. Kidney biopsies are invasive and too costly to assess the course of LN. The objective of this study is to develop and initially validate for Children a Renal Activity Index (C-RAI) to non-invasively monitor LN activity, considering both traditional measures of LN (LN-TM) and recently discovered renal biomarkers (RBM).

Methods In this ongoing prospective study, 83 children with LN were studied at the time of the kidney biopsy; LN-TM [GFR, complements, anti-dsDNA antibodies, urinary protein/creatinine ratio], clinical indices [Systemic Lupus International Collaborating Clinics Renal Activity Score (SLICC-RAS), renal domain score of BILAG (BILAG-R) and SLEDAI (SLEDAI-R)] were all obtained, and the RBM (Table 1) were measured. Histological findings were rated by a single nephropathologist who provided ISN/RPS class, NIH Glomerular Activity Index (GLAI; range 0-24) and Tubulointerstitial Activity Index (TIAI; range = 0-21) scores (Criterion Standards). Prior to statistical analysis, RBM levels were normalized by urine creatinine and logarithmically transformed. LN-TM, RBM and clinical indices that showed significance in univariate analysis at a p-value<0.10 were considered in stepwise multivariate logistical regression models as C-RAI candidate predictors, using the GLAI and TIAI as dependent variable (outcome). The accuracy of the C-RAI of predicting and discriminating LN activity was assessed by receiver-operating characteristic curve (ROC) analysis.

Results Means and percentages of the values of LN-TM, clinical indices and RBM levels are summarized inTable1. Based on multivariate logistical regression modeling, histological activity measurement does not necessitate consideration of clinical indices but rather select LN-TM and RBM. Levels of C3, NGAL, CP, MCP1 and TF were found to be  candidate C-RAI’s for predicting high LN activity (GLAI>10) with outstanding  accuracy [area under the ROC curve (AUC) = 0.9]. NGAL and HPX were excellent predictors of high interstitial inflammation with active LN (TIAI > 5; AUC = 0.88) (Figure1).

Conclusion C3 level, NGAL, CP, MCP1, TF, and HPX are good potential components for C-RAI to measure histological LN activity in the glomeruli and interstitium. Confirmation in a larger data set is required.