Anti-neutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) is a small-vessel necrotizing vasculitis that frequently affects the kidneys, manifesting as a pauci-immune focal and segmental necrotizing and crescentic glomerulonephritis.1 Renal involvement remains a major determinant of adverse kidney and patient outcomes,2,3 with approximately 20–40% of patients progressing to end-stage kidney disease (ESKD).4–6 Treatment of AAV relies on intensive immunosuppression to rapidly control inflammation and prevent irreversible organ damage. However, therapy adverse effects are considerable and contribute to significant patient morbidity and mortality, with infections leading the cause of early mortality in the first year of diagnosis.7 Therefore, identifying accurate predictors of kidney outcome at the time of diagnosis remains a clinical priority to guide treatment decisions.

Several clinical and histopathological factors have been associated with long-term kidney outcomes, including older age, baseline kidney function, hypertension, and relapses,4,6,8–10 as well as biopsy findings of percentage of normal glomeruli9,10 and chronic lesions like interstitial fibrosis and tubular atrophy (IFTA).11 To standardize risk assessment, different scoring systems have been developed to inform about the risk of ESKD in AAV (Fig. 1).

Fig. 1.

ANCA-associated vasculitis kidney histological scores.

The European Vasculitis Study Group (EUVAS) histological classification12 stratifies kidney biopsy findings into focal, crescentic, mixed, and sclerotic classes according to the predominant glomerular lesion. While the original cohort demonstrated a progressive decline in kidney prognosis from focal to sclerotic classes, subsequent validation studies highlighted its limited discriminative ability, particularly between crescentic and mixed classes, and underscored the importance of incorporating tubulointerstitial lesions to improve prognostic accuracy.9,13

Later, the Mayo Clinic Chronicity Score (MCCS) was proposed by Sethi et al.14 to standardize the reporting of chronic histological changes in glomerulonephritis and was subsequently validated for AAV.11 Higher chronicity grades have been consistently associated with lower rates of renal recovery, increased risk of ESKD, and mortality,11 establishing the MCCS as a valuable prognostic tool in AAV.4,6,15

In turn, Brix et al.16 introduced the clinicopathological ANCA Renal Risk Score (RRS), which combined the percentage of normal glomeruli, the degree of IFTA, and the baseline glomerular filtration rate (eGFR). It demonstrated excellent discriminative ability in the development cohort and has since been validated in multiple independent studies.4–6,15,17 More recently, the RRS was revised into the Improved ANCA Kidney Risk Score (AKRiS).18 By replacing eGFR with serum creatinine, reweighting the contribution of chronic lesions, and introducing a fourth risk category, AKRiS refined the original model and demonstrated excellent performance for predicting ESKD in the derivation cohort and subsequent validation studies.15,18,19

Although these scores have been validated individually, direct comparisons between classification systems, especially the most recent AKRiS, remain scarce. Furthermore, their performance under current treatment regimens has not been fully established, and the extent to which they provide complementary prognostic information remains underexplored. Hence, we aimed to evaluate and compare the performance of current histological and clinicopathological scores in predicting ESKD in a contemporary cohort of patients with AAV and explore their complementary role in enhancing risk stratification.

In this study, we reinforced the role of chronic tubulointerstitial lesions and MCCS for predicting AAV kidney outcomes independent of kidney function and induction therapy and confirmed that incorporating kidney function at the time of biopsy in RRS and AKRiS enhanced risk stratification. The AKRiS differentiated kidney survival among all risk categories and showed excellent discriminative performance for predicting ESKD in AAV, which remained robust after internal bootstrap validation. Cross-comparison analysis revealed a significant concordance between risk groups of MCCS, RRS, and AKRiS and suggested a complementary role for these scores in refining risk stratification, particularly among patients with intermediate risk.

The prognostic value of the Berden Classification has been previously established.12 Focal and sclerotic classes were typically associated with the best and worst kidney survival, respectively, while crescentic and mixed classes exhibited intermediate prognosis.13,21,22 In our cohort, kidney survival differed significantly only between the sclerotic and the remaining histological classes. Earlier diagnosis and prompt initiation of immunosuppressive therapy have likely improved outcomes in patients with active or potentially reversible lesions, such as those classified within crescentic and mixed patterns, justifying the absence of significant survival differences among these groups in contemporary cohorts,15,23 when compared to historical ones.12,13 These findings reinforce recent observations that Berden glomerular-based classification alone may lack sufficient discriminative power for reliable risk stratification in current AAV patients.9,10,15

The MCCS has been previously associated with lower rates of kidney recovery11 and a higher risk of ESKD with increasing chronicity grades in AAV.6,11,24 In our cohort, higher degrees of chronicity were associated with increasing risk of ESKD, with statistically significant differences in kidney survival between risk groups. Furthermore, MCCS remained an independent predictor of ESKD after adjustment for serum creatinine, KRT at presentation, plasma exchange and immunosuppressant induction therapy, and demonstrated strong discriminative ability (optimism-corrected C-index 0.821 [95% CI 0.730–0.931]), reinforcing its prognostic value in AAV.6,15,24 The moderate concordance observed between MCCS and both RRS and AKRiS at the low- and high-risk ends further supports its validity for kidney prognosis in AAV. Additionally, the variability observed in intermediate-risk groups suggests that the scores offer distinctive prognostic information and highlight the complementary role of combining MCCS with the more recent clinicohistopathological scores to refine risk stratification in intermediate prognosis subgroups.

The RRS proposed by Brix et al.,16 integrating the percentage of normal glomeruli and chronic tubulointerstitial damage with baseline serum eGFR, demonstrated good discriminative performance, with an optimism-corrected C-index of 0.779 [95% CI 0.641–0.957] after internal validation, aligning with previous RRS validation studies, ranging from 0.696 by Stella et al.,6 to 0.779 in the RRS development cohort16 and 0.877 in a Scottish National Cohort.5 However, Kaplan–Meier analysis only distinguished renal survival between high and low-risk groups. Sandino-Bermúdez et al.15 also noted that the RRS distinguished high-risk from other categories but did not differentiate between the low and medium risk groups. These findings suggest that the RRS medium-risk category may capture a more heterogeneous patient population with a wide range of risk profiles. One explanation lies in the use of eGFR, which condenses a broad range of creatinine values into narrow eGFR intervals,18 reducing its reliability in acute kidney injury. Additionally, given that IFTA emerged as the strongest independent predictor of ESKD in our cohort, the IFTA binary classification in RRS, when compared to the more granular grading of MCCS and reweighted categories in AKRiS, may have contributed to its lower discriminative performance, particularly among intermediate-risk patients.

In turn, by replacing eGFR with serum creatinine and creating an additional risk category, the revised AKRiS allowed for better risk discrimination and model performance.18 In this study, it was a predictor of ESKD independent of the KRT, plasma exchange and immunosuppressive treatment, differentiated renal survival among all risk categories, and provided excellent discriminative performance for predicting ESKD (optimism-corrected C-index 0.871 [95% CI 0.781–0.989]), exceeding the results of previous validation cohorts.15,19 Despite the strong association between RRS and AKRiS categories, we found that RRS ‘High’ risk patients were distributed over the AKRiS ‘Medium’, ‘High’, and ‘Very High’, also demonstrating the refined discriminative ability of AKRiS among higher risk patients and possibly justifying the survival differences between scores. Interestingly, all patients in the ‘Very High’ risk group started KRT months after biopsy, reflecting the patients with chronic irreversible damage who would probably not benefit from immunosuppressive therapy. Although the small number of patients in this risk category limits the generalizability of these results, it further supports the improved prognostic role of AKRiS.

One key finding of our study was the prognostic impact of chronic structural damage in predicting ESKD and its central role in the performance of MCCS, RRS, and AKRiS. Chronic irreversible changes are less influenced by therapy and inform about the potential for renal recovery,3,11 providing a reliable tool for predicting long-term kidney outcomes and guiding treatment decisions at the time of diagnosis.25 Although other authors have shown that the inclusion of post-treatment parameters, such as eGFR recovery,8,15 proteinuria,26,27 and kidney relapses8 improves the scores long-term prognostic performance, chronic damage remains one of the main predictors to guide treatment decisions at the time of diagnosis.28 Additionally, MCCS, RRS and AKRiS remained independently associated with ESKD after adjustment for KRT requirement, plasma exchange and immunosuppressive induction therapy, further reinforcing their role in treatment decision-making.

Emerging evidence suggests that molecular profiling of diagnostic kidney biopsies may further refine prognostic assessment in AAV-GN. Transcriptomic signatures appear capable of capturing biologic heterogeneity and active pathogenic pathways that are not fully reflected by semi-quantitative structural scoring, potentially enhancing prediction of kidney survival, particularly at mid- and long-term follow-up.29 These findings highlight the complexity of renal injury mechanisms in AAV and suggest that future prognostic frameworks may benefit from integrating structural and molecular information to achieve more individualized risk stratification.

We acknowledge certain limitations to this study. First, its single-centre and retrospective design introduce a potential for selection and information bias, which may limit the generalizability of our findings. Second, the relatively small sample size and limited number of ESKD events constrained the power of the survival analyses and the number of covariates that could be included in multivariable models. Third, patients diagnosed in the most recent years had substantially shorter follow-up compared with those included earlier. This may lead to underestimation of long-term ESKD risk and influence survival curve interpretation beyond the first year after biopsy. Therefore, longer and more homogeneous follow-up in contemporary cohorts will be important to confirm the long-term performance of these scores. Fourth, our cohort did not include patients treated with avacopan, a C5a receptor inhibitor increasingly incorporated into AAV induction regimens. As avacopan may influence renal recovery and long-term kidney outcomes, the performance of conventional histological and clinicopathological scores should be reassessed in cohorts managed under contemporary glucocorticoid-sparing strategies.

Despite these limitations, this study presents a contemporary cohort of patients with AAV, integrating multiple validated histological scoring systems, including the recently proposed AKRiS. It provides novel insights into score performance and inter-score association, offering clinically relevant data for risk stratification at diagnosis. Although the sample size was limited, patients were reasonably distributed across risk categories, allowing meaningful comparative analyses, and internal bootstrap validation supported the stability of the discriminative performance. Furthermore, most patients received induction immunosuppression and, in contrast to previous validation cohorts,5,6,15,24 a higher proportion of patients were treated with rituximab or combined regimens of rituximab and cyclophosphamide, supporting the relevance of these scores under current standards of care.

In conclusion, our study reinforces the value of kidney biopsy in AAV and validates the clinical utility of current histological scoring systems in a contemporary cohort. The more recent AKRiS showed better discriminative ability, differentiating risk across all categories. The complementary use of histological and clinicopathological scores in intermediate-risk groups may enhance stratification at diagnosis and support more informed therapeutic decisions.

This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Ethics Committee of Unidade Local de Saúde de Braga (No. 255_2025).

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

The authors have no relevant financial or non-financial interests to disclose.

The data underlying this article will be shared on reasonable request to the corresponding author.