Predictive Factors of Renal Recovery and Progression to End-Stage Kidney Disease in Patients With Antineutrophil Cytoplasmic Autoantibody-Associated Vasculitis With Severe Kidney Disease

Introduction A significant number of patients with antineutrophil cytoplasmic antibodies (ANCA)- associated vasculitis (AAV) with glomerulonephritis (AAV-GN) still progress to end-stage kidney disease (ESKD, estimated glomerular filtration rate [eGFR] <15 ml/min per 1.73 m2) despite advances in remission-induction treatment. Methods This is a retrospective cohort study on myeloperoxidase (MPO)-ANCA or proteinase 3 (PR3)-ANCA positive patients with AAV (microscopic polyangiitis, MPA; or granulomatosis with polyangiitis, GPA) and eGFR <15 ml/min per 1.73 m2 or ESKD at presentation. Renal recovery, dialysis discontinuation, and persistence of ESKD after standard remission-induction, with or without the use of plasma exchange (PLEX) were analyzed. Results We analyzed 166 patients with biopsy-proven active AAV-GN and eGFR <15 ml/min per 1.73 m2 at the time of diagnosis. Patients received glucocorticoids with cyclophosphamide (CYC) (n = 84) or with rituximab (RTX) (n = 72) for remission-induction, and 49 received PLEX. The predictors of renal recovery were erythrocyte sedimentation rate, serum creatinine (SCr) at diagnosis, and minimal or mild chronicity changes. We further analyzed 71 patients who started dialysis with or without PLEX within 4 weeks of AAV-GN diagnosis. The predictors of dialysis discontinuation were minimal chronicity changes in kidney biopsy at diagnosis (odds ratio = 6.138; 95% confidence interval [CI]: 1.389–27.118; P = 0.017). Predictors of persistence of ESKD within 12 months included higher SCr at diagnosis (incidence rate ratio [IRR] = 1.086; 95% CI: 1.005–1.173; P = 0.037), and moderate (IRR = 3.797; 95% CI: 1.090–13.225; P = 0.036), or severe chronicity changes in kidney biopsy (IRR = 5.883; 95% CI: 1.542–22.439; P =0.009). Conclusion In our cohort, kidney recovery, dialysis discontinuation, and persistence of ESKD in patients with AAV-GN and eGFR <15 ml/min per 1.73 m2 depended on SCr and histologic findings on kidney biopsies at the time of diagnosis and was not affected by the addition of PLEX.

chronicity changes on biopsies were the most important predictive factors for renal recovery, even when adjusted for treatment with PLEX (Supplementary Table S5).

ESKD within 12 months (n=150) Multivariable logistic regression analysis
showed that higher SCr, lower hemoglobin at diagnosis and moderate to severe chronicity grading on kidney biopsy were the most important predictive factors for progression to ESKD at 12 months, even when adjusted for the treatment with PLEX (Supplementary Table S6).

Transient versus persistent need for dialysis (n=71).
Multivariable logistic regression analysis showed that minimal chronicity changes or focal involvement were the most important predictors of transient dialysis, even when adjusted for the treatment with PLEX (Supplementary Table S7).
When comparing CYC vs. RTX, we excluded from the analysis the 6 patients from the Vall d'Hebron cohort that have received both CYC and RTX.There were no differences on the BVAS/WG scores, frequency of alveolar hemorrhage between groups, and SCr and eGFR at diagnosis.A higher proportion of patients scored as moderate in the chronicity score in the RTX group (55.1% vs. 39.5%,p=0.024).PLEX was added to standard immunosuppressive remission-induction therapy more frequently in patients who received CYC (42.0% vs. 15.9%,p=0.001)whereas IV methylprednisolone was more commonly added in patients treated with RTX (91.3% vs. 67.9%,p<0.001).Patients who received CYC relapsed more (33.3%vs. 15.9%,p=0.015).There were no differences between groups in the frequency of renal recovery, dialysis (permanent or transient), ESKD at 12 months or infections.Patients treated with CYC had worse survival, 37 (45.7%)died when compared with 17 (24.6%) in the RTX group (p=0.007).In the decade analysis, treatment choices reflected the standard of care with the incremental use of RTX and discontinuation of addition of PLEX in the more recent years (Supplementary Table S10).Importantly, renal recovery and survival improved along the years whereas the ESKD frequency remained constant between decades, highlighting this unmet need.Abbreviations: AAVanti-neutrophil cytoplasmic antibody associated vasculitis; eGFRestimated glomerular filtration rate; FUfollow-up; IQRinterquartile range; KFkidney failure; n -number.

Table S5 -
Univariable and Multivariable logistic regression of the predictive factors for renal recovery at 12 months (n=145).
Supplementary TableS6-Univariable and Multivariable logistic regression of the predictive factors for ESKD at 12 months (n=166).

Table S8 -
Clinical characteristics and Outcomes with Anti-NeutrophilCytoplasmic Antibodies (ANCA) associated vasculitis (AAV) patients with glomerulonephritis (AAV-GN) and eGFR < 15 mL/min/1.73m 2 at diagnosis who started dialysis in the first 4 weeks of AAV-GN diagnosis according with PLEX status (n=71).