Validation and modification of renal risk score for prediction of long-term kidney survival in patients with MPO-ANCA-associated glomerulonephritis.

BACKGROUND

Anti-neutrophil cytoplasmic antibody (ANCA) Renal Risk Score has not been widely validated in Chinese patients with myeloperoxidase -ANCA-associated glomerulonephritis and its predictive ability needs to be improved.

METHODS

Three hundred and forty patients with biopsy-proven myeloperoxidase-ANCA-associated glomerulonephritis were included in this study. They were divided into an oliguric group (urine volume < 400 ml/24 h) and a non-oliguric group (urine volume ≥ 400 ml/24 h). The ANCA Renal Risk Score and Berden classes were used to predict the risk of end-stage kidney disease (ESKD), and Cox regression analysis was performed to evaluate the impact of oliguria on ESKD risk.

RESULTS

The predictive performance of ANCA Renal Risk Score was significantly higher than that of Berden classes (AUC: 0.79 vs. 0.709, P = 0.003). Thirty-six (10.6%) patients presented with oliguria. Patients in the oliguric group had significantly lower levels of baseline estimated glomerular filtration rate (eGFR) [6.51(4.55-7.72) vs. 21.22(11.49-36.63) ml/min/1.73 m, P < 0.001], hemoglobin (78.33 ± 16.75 vs. 92.47 ± 18.95 g/L, P < 0.001), serum albumin (32.01 ± 5.92 vs. 36.22 ± 4.84 g/L, P < 0.001), and a lower percentage of normal glomeruli [6.86(0-17.39)% vs. 18.18(9.09-35)%, P < 0.001]. Consistently, the oliguric group had a higher percentage of patients that progressed to ESKD (83.3% vs. 36.2%, P < 0.001). Multivariate Cox regression analysis showed that oliguria was an independent risk factor for ESKD [HR = 2.38(1.39-4.06), P = 0.002]. Oliguria scores (3 points for presence and 0 for absence) were incorporated into the ANCA Renal Risk Score, resulting in ANCA Renal Risk Score-U. The patients were then categorized into four risk groups: low-(0-2 points), moderate-(3-7 points), high-(8-11 points) and very high-(12-14 points). The incidences of ESKD in these groups were 11.1%, 30%, 72.2% and 95.8%, respectively. The predictive efficacy of ANCA Renal Risk Score-U in predicting ESKD risk was significantly higher than that of the ANCA Renal Risk Score (AUC: 0.812 vs. 0.79, P = 0.025).

CONCLUSIONS

This study has validated the ANCA Renal Risk Score for predicting kidney outcomes among Chinese patients with myeloperoxidase-ANCA-associated glomerulonephritis. Furthermore, the ANCA Renal Risk Score-U model with oliguria as a variable can further improve the prediction of ESKD risk in patients with myeloperoxidase-ANCA-associated glomerulonephritis.