Prediction of relapses in patients with small vessel vasculitides: a multicenter cohort study on histopathological risk patterns

Abstract

Management of ANCA-associated vasculitis (AAV) has significantly improved, yet up to 40% of patients experience relapses, leading to worse long-term outcomes and organ damage. This multicenter cohort study investigated the prognostic value of histopathological patterns in renal AAV for predicting relapse risk and treatment response. We retrospectively analyzed 264 patients with newly diagnosed granulomatosis with polyangiitis (GPA) or microscopic polyangiitis (MPA) with renal involvement recruited at four tertiary rheumatology and/or nephrology centers. Baseline clinical data, disease activity scores, and kidney biopsy findings were assessed. Patients were followed for at least twelve months. An overall relapse was defined as renewed disease activity requiring changes in maintenance therapy, initiation of a new induction regimen, or an increase in prednisone dosage > 10 mg/day. A severe relapse required new immunosuppressive induction therapy. After a median follow-up of twelve months the renal domain of the Birmingham Vasculitis Activity Score (BVAS) and its subcategories were evaluated. AAV patients with moderate-to-high (> 25%) interstitial fibrosis and tubular atrophy (IFTA) had a lower risk for both, overall and severe relapses. Severe relapses occurred more frequently in older and male patients. Furthermore, we identified glomerular sclerosis (≥ 50%) to be the strongest predictor for ongoing activity in the renal BVAS domain. Histopathological features do not only help to predict renal recovery and need for dialysis but also to forecast relapse risk and renal BVAS activity. Incorporating these patterns into clinical decision-making could enable more personalized therapy approaches, highlighting the need for prospective validation.

Keywords: Granulomatosis with polyangiitis; Kidney diseases; Microscopic polyangiitis; Recurrence; Risk factors; Vasculitis.

Fig. 1

Change of ANCA titers during relapse. ANCA Anti-neutrophil cytoplasmic antibodies

Fig. 2

Histopathological patterns from kidney samples in relapse and non-relapse patients. Line indicating median, dot indicates mean. IFTA interstitial fibrosis and tubular atrophy

Fig. 3

Forest plot displays hazard ratio for different predictors in a multivariate Cox regression model for the event overall relapse. Predictors included were significant variables from univariate analyses, plus key covariates: age, male gender, GPA subtype, baseline eGFR, baseline BVAS score, and RTX induction treatment. Model selection utilized stepwise backward/forward selection based on Akaike information criterion (AIC) to optimize predictive performance. Model’s discrimination was moderate, with a concordance index (C-index) of 0.59 (standard error = 0.037). Overall model fit was statistically significant as indicated by the likelihood ratio test (p = 0.01). GPA granulomatosis with polyangiitis, eGFR estimated glomerula filtration rate, BVAS Birmingham Vasculitis Activity Score, RTX rituximab, IFTA interstitial fibrosis and tubular atrophy

Fig. 4

Forest plot illustrating the hazard ratios (HR) for different predictors in a multivariate Cox regression model for the event severe relapse. Predictors included were significant variables from univariate analyses, plus key covariates: age, male gender, GPA subtype, baseline eGFR, baseline BVAS score, and RTX induction treatment. Model selection utilized stepwise backward/forward selection based on Akaike information criterion (AIC) to optimize predictive performance. The model showed a concordance index (C-index) of 0.73 (SE = 0.048), indicating good discrimination ability. Overall model fit was statistically significant, as demonstrated by the likelihood ratio test (p < 0.001). GPA granulomatosis with polyangiitis, eGFR estimated glomerula filtration rate, BVAS Birmingham Vasculitis Activity Score, RTX rituximab, IFTA interstitial fibrosis and tubular atrophy