Development and validation of a renal risk score in ANCA-associated glomerulonephritis

Abstract

Predicting renal outcome in antineutrophil cytoplasmic antibody (ANCA)-associated glomerulonephritis (GN) remains a major challenge. We aimed to identify reliable predictors of end-stage renal disease (ESRD) and to develop and validate a clinicopathologic score to predict renal outcome in ANCA-associated GN. In a prospective training cohort of 115 patients, the percentage of normal glomeruli (without scarring, crescents, or necrosis within the tuft) was the strongest independent predictor of death-censored ESRD. Regression tree analysis identified predictive cutoff values for three parameters: percentage normal glomeruli (N0 >25%, N1 10 to 25%, N2 <10%), percentage tubular atrophy and interstitial fibrosis (T0 ≤25%, T1 >25%), and estimated glomerular filtration rate at the time of diagnosis (G0 >15 ml/min/1.73 m2, G1 ≤15 ml/min/1.73 m2). Cox regression analysis was used to assign points to each parameter (N1 = 4, N2 = 6, T1 = 2, G1 = 3 points), and the resulting risk score was used to classify predicted ESRD risk as low (0), intermediate (2 to 7), or high (8 to 11 points). The risk score accurately predicted ESRD at 36 months in the training cohort (0%, 26%, and 68%, respectively) and in an independent validation cohort of 90 patients (0%, 27%, and 78%, respectively). Here, we propose a clinically applicable renal risk score for ANCA-associated GN that highlights the importance of unaffected glomeruli as a predictor of renal outcome and allows early risk prediction of ESRD.