Abstract
OBJECTIVE: Fixed and tailored rituximab retreatment strategies to maintain remission in antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) are associated with tradeoffs. The current study was undertaken to develop a simulation model (AAV-Sim) to project clinical outcomes with these strategies. METHODS: We developed the AAV-Sim, a microsimulation model of clinical events among individuals with AAV initiating treatment to maintain remission. Individuals transition between health states of remission or relapse and are at risk for severe infection, end-stage renal disease, or death. We estimated transition rates from published literature, stratified by individual-level characteristics. We performed validation using the mean average percent error (MAPE) and the coefficient of variation of root mean square error (CV-RMSE). In internal validation, we compared model-projected outcomes over 28 months with outcomes observed in the Rituximab versus Azathioprine in ANCA-Associated Vasculitis 2 (MAINRITSAN2) trial, which compared fixed versus tailored retreatment. In external validation, we compared outcomes with fixed rituximab retreatment from the AAV-Sim to outcomes from the MAINRITSAN1 trial and an observational study. RESULTS: The AAV-Sim projected outcomes similar to those in the MAINRITSAN2 trial, including minor (AAV-Sim 6.0% fixed versus 7.3% tailored; MAINRITSAN2 6.2% versus 8.6%; MAPE 3% and 15%) and major relapse (AAV-Sim 3.5% versus 5.5%; MAINRITSAN2 3.7% versus 7.4%; MAPE 5% and 26%), severe infection (AAV-Sim 19.4% versus 11.1%; MAINRITSAN2 19.8% versus 10.2%; MAPE 2% and 9%), and relapse-free survival (AAV-Sim 84.8% versus 82.3%; MAINRITSAN2 86% versus 84%; CV-RMSE 2.3% and 2.5%). Similar performance was observed in external validation. CONCLUSION: The AAV-Sim projected a range of clinical outcomes for different treatment approaches that were validated against published data. The AAV-Sim has the potential to inform management guidelines and research priorities.