Op0142 Increased Risk of Severe Infection in Patients With Antineutrophil Cytoplasmic Antibody-Associated Vasculitides: A Population-Based Trend Analysis

Background: Patients with ANCA-associated vasculitis (AAV) exhibit higher morbidity and premature mortality rates compared to the general population, primarily due to extensive inflammation and the side effects of treatment regimens like high-dose glucocorticoids and immunosuppressive agents. Prior research indicates a sixfold increased infection risk in AAV patients, but this risk's temporal evolution remains unexplored. Moreover, existing studies may not accurately reflect the infection risk due to limitations such as hospital-based samples, small cohorts, differing AAV definitions, and incomplete adjustment for confounders. Objectives: We: a) quantified the independent risk of the first severe infection in AAV patients; b) tracked the infection risk trajectory in AAV patients; and c) compared the rate of recurrent severe infections in AAV patients versus matched controls. Methods: We executed a 1:10 matched cohort study, comparing incident AAV patients with age-, sex- matched individuals from the general population using administrative health data. The primary outcome was the first post-AAV severe infections necessitating admission to hospital or occurring during a hospitalization. We employed multivariate Cox models for infection onset timing. The secondary outcome involved longitudinal severe infection counts, analyzed using a two-part zero-inflated Poisson mixed model (ZIPMM). Results: In our cohort of 549 newly diagnosed AAV patients (58% female, average age 54.3 years), we observed that one in three developed severe infections. This represented a 3.8-fold increase in infection risk compared to matched non-AAV controls during follow-up. AAV patients had a higher rate of severe infections than the general population throughout the whole study period (Figure 1). Our statistical analysis (Table 1), highlighted a two-phase reduction in infection risk among AAV patients: initially, there was a sharp decline, followed by a more gradual decrease. Specifically, at the time of diagnosis, AAV patients had a substantially higher risk of developing an severe infection (OR=77.02, 95% CI: 11.67, 508.6) compared to controls. Additionally, the rate of severe infections at diagnosis was also higher in the AAV group (RR=2.55, 95% CI: 1.36, 4.78) compared to controls. Post-diagnosis, the infection risk trend shifted, particularly after two years. The odds ratio for infection incidence in AAV patients decreased to 0.32 (95% CI: 0.16, 0.62) before the end of the second year, indicating a lowering risk of developing severe infection over time. However, after the second year, the decrease in risk slowed down, with an odds ratio of 0.85 (95% CI: 0.76, 0.96), suggesting a less pronounced yet still significant reduction in risk of developing severe infection. The initial stable rate of infections in the first two years (RR=0.97, 95% CI: 0.79, 1.20) was followed by an evident increase thereafter (RR=1.11, 95% CI: 1.04, 1.17) in AAV patients. Conclusion: This comprehensive study, which is the largest of its kind on incident AAV to date, is the first to accurately measure the risk of severe infections in AAV patients using a large, general population-based cohort. Our findings, through ZIPMM model, show a clearer picture of how the risk of infection in AAV patients changes over time compared to the general population. We observed that the decreasing gap in infection risk between AAV patients and the general population over time was mainly because fewer AAV patients were at risk for severe infections as time went on, not necessarily because those at risk were having fewer repeated infections. Understanding who is most at risk for infections and repeated infections can help us use healthcare resources more effectively and tailor our treatment strategies. Further studies are needed to explore how inflammation, medications contribute to infections in AAV patients. Table 1. Secular trend of infection count in AAV relative to the general population using ZIPMM model REFERENCES: NIL. Acknowledgements: NIL. Disclosure of Interests: None declared.