1080: Performance of Risk Prediction Models for Relapse and Severe Infection at the End of 18-month Rituximab Maintenance Therapy in ANCA-associated Vasculitides: Data from the MAINRITSAN Trials
Florence Delestre1, Pierre Charles2, Loïc Guillevin3, Raphaël Porcher4 and Benjamin Terrier3, 1cochin Hospital, Paris, France, 2Institut Mutualiste Montsouris, Paris, France, 3National Referral Center for Rare Systemic Autoimmune Diseases, Cochin Hospital, Paris, France, 4Université Paris Cité, Hôtel-Dieu, Paris, France
Background/Purpose: Long-term follow-up analysis of the three MAINRITSAN trials confirmed the efficacy of rituximab (RTX) in preventing relapses in ANCA-associated vasculitis (AAV), especially RTX administration every 6 months for 18 months. Extending RTX maintenance therapy beyond the first 18 months of maintenance therapy may be necessary for some patients. We aimed to analyze the performance of relapse and infection risk prediction models to help guide decision making regarding the extension of maintenance treatment with RTX.
Methods: The MAINRITSAN randomized controlled trials compared fixed 500 mg RTX infusion at D0, D14, M6, M12, and M18 with other regimens. All patients enrolled in these trials were prospectively followed until December 2020 and their data were pooled to analyze relapses and adverse events. We performed external validation of the relapse and infection risk prediction models reported by McClure et al. The primary efficacy endpoint was the time from RTX termination to first relapse. The primary safety endpoint was the time from RTX termination to the first severe infection.
Results: Two hundred and seventeen patients treated with RTX for 18 months were included in this analysis, including 56 patients from MAINRITSAN1 trial and 161 from MAINRITSAN2 trial. Median follow-up from the last RTX infusion was 63 months (IQR 16-80). Seventy-six patients relapsed, including 46 with a major relapse, corresponding to an overall relapse-free survival at 72 months of 55% (95% CI 48-64). Severe infections were observed in 33 patients, corresponding to a cumulative incidence of severe infection at 72 months of 20%.
The relapse prediction model, based on male gender, age >60 years, ANCA-positive at last RTX infusion, relapsing disease, ENT involvement and prednisone dose at last visit, tended to be associated with the risk of overall relapse (coefficient 0.81, p=0.096) and was significantly associated with the risk of major relapse (coefficient 1.34, p=0.04). It defined patients at high and low risk of relapse, those at high risk of relapse having shorter overall relapse-free survival than those at low risk (HR 1.47, 95% CI 0.90-2.39, p=0.12), with a more pronounced difference for major relapse-free survival (HR 1.91, 0.97-3.76, p=0.059).
According to ANCA status, overall relapse-free survival was poorer in PR3-ANCA patients that in MPO-ANCA, without any difference according to relapse score. In contrast, the relapse score was highly predictive of major relapse in PR3-ANCA patients (coefficient 2.76, p=0.01) but not in MPO-ANCA patients.
The infection prediction model, based on male gender, structural lung disease, diabetes, occurrence of infections during maintenance therapy and gammaglobulin level, identified patients at high risk for infection showing a shorter severe infection-free survival than low-risk patients (HR 2.93, 95% CI 1.28-6.72, p=0.011).
Conclusion: We validate here risk prediction models for relapse and severe infection at the end of 18-month RTX maintenance therapy in AAV. The relapse score is highly predictive of major relapse in PR3-ANCA patients and could help physicians in decision-making regarding whether to extend RTX maintenance therapy in PR3-ANCA.
Disclosures: F. Delestre, None; P. Charles, None; L. Guillevin, Roche; R. Porcher, None; B. Terrier, AstraZeneca, GlaxoSmithKline, Bristol-Myers Squibb(BMS), Eli Lilly, LFB, Boehinger Ingelheim, Vifor Pharma, Pfizer, Roche.
