Background/Purpose: Evaluation of infection risk in ANCA-associated vasculitis (AAV) has been limited to small, selected populations and/or serious episodes. In this large study, we aimed to track the risk of all infections in AAV and compare it to the general population.

Methods: A longitudinal matched-cohort study was developed using Scottish administrative health-data registries. Uniquely, these resources provide over 97% population coverage and offer the capacity to link microbiological laboratory data from all levels of health care in Scotland. AAV patients fulfilling European Medicines Agency criteria were identified by clinicians across Scotland. Each was matched with up to 5 general population controls by age (±2 years), sex and geography. Both cohorts were followed from the date of AAV diagnosis (same day assigned for matched controls) until death or 02/28/17, whichever came first. Data on all infections were retrieved from the Electronic Communication of Surveillance in Scotland Database. Descriptive statistics were used to compare infection types in both cohorts. Incidence rate ratios (IRR) and 95% confidence intervals (CI) were computed using multilevel Poisson regression.

Results: A total of 379 AAV patients (51.7% male; median age: 61.6 years) and 1859 general population controls were followed for a median of 3.5 years. During follow-up, AAV patients and controls developed 1127 and 1256 infections, respectively (55.7% of AAV, 15.8% of controls; IRR 6.7, 95%CI 5.2-8.6, p<0.0001). Over 60 infection types were identified in AAV patients in whom Escherichia coli was the most commonly observed infection (15.8% vs 5.2%, p<0.001). However, compared to the general population, AAV patients were at greatest risk of contracting Rhinovirus (IRR 9.5, 95% CI 3.5-25.8, p<0.0001), Mycobacterium (IRR 9.5, 95% CI 2.5-34.1, p<0.0001), Enterobacter (IRR 8.8, 95% CI 3.0-25.7, p<0.0001), and Citrobacter (IRR 6.4, 95% CI 2.0-20.4, p<0.0001), respectively. Interestingly, the highest rates of infections in AAV were observed during the first two years of AAV diagnosis (p<0.0001, Figure 1). Although this risk decreased over time, AAV patients continued to be at a higher risk of infection than controls.

Conclusion: To our knowledge, this is the first and the most comprehensive study to analyse both the risk and type of infections using laboratory records and to track infection risk over time in AAV. Our findings indicate that AAV patients face a high risk of infections, especially in the first two years after AAV diagnosis. This risk reduces with time, but remains significantly greater than that in the general population even after eight years of follow-up.

Figure 1. Comparison of the rate of infections in ANCA-associated vasculitis and the general population