Abstract Background and Aims Infection due to the immunosuppressive treatment, is among the most frequent causes of hospitalization and death in patients with ANCA-associated vasculitis (AAV). Information on prevailing causal microbial aetiologies is crucial for improving treatment and prevention of such events. Here we examined causal bacteriology, incidence, and risk of blood stream infections (BSIs) in incident and prevalent AAV patients, as compared to the general population. Method All data were retrieved from the Danish nationwide registries and the Danish Microbiology Database. Prevalent patients between 1998-2010 were all included at the same index date 1.1.2010; Incident patients diagnosed after 2010 were included with index being the day of diagnosis. Both groups were followed until first time BSI, death, or a maximum of 12 months. Background controls were matched 1:4 on age and sex. Cumulative incidence was assessed by the Aalen-Johansen estimator, and cox analyses adjusted for age, sex, kidney involvement, hypertension, diabetes, dialysis, and plasma exchange (PLEX), were used to model survival time. Results A total 19 (2.3%) BSIs were identified in 818 prevalent patients, and 55 (5.9%) BSIs in 939 incident patients (P < 0.001). As in the background control group, primary causal microbial agents among prevalent patients were E. coli (26.3%) and S. aureus (21.1%), with an overall Gram-negative predominance (58.8%). E. coli (17.5%) remained the most frequent microbial cause of BSIs in incident patients, however with a relatively higher frequency of gram-positive bacteria (50.9%) dominated by Coagulase-negative Staphylococci (12.3%), Enterococci (14%), and Streptococci (14%). Difference in bacteriology between prevalent and incident patients disappeared when patients on dialysis were excluded, whereas the difference in BSI frequency remained significant (p < 0.001). One-year HR of first-time BSI was significantly higher in patients with AAV as compared to controls (prevalent AAV: HR 3.17 [95% CI 1.63-6.18], p < 0.0001; Incident AAV; HR 8.64 [95% CI 4.88-15.31], p < 0.0001), with significant difference between incident and prevalent patients (HR 2.73 [95% CI 1.39-5.35], p = 0.003). Dialysis (HR 3.15 [95% CI 1.86-5.31], p < 0.0001) and PLEX (HR 1.87 [95% CI 1.05-3.33], p < 0.033) were solitary risk factors of BSI as well as higher age, hypertension, diabetes, and kidney disease (Figure). Risk of BSI remained significantly increased for both prevalent and incident patients as compared to matched background controls when patients on dialysis were excluded. Conclusion E. coli and S. aureus were predominant causal isolates in prevalent AAV patients and controls, and the gram-positive predominance seen with incident AAV was primarily caused by patients on dialysis. Risk of BSI was higher in incident AAV as compared to prevalent AAV, and the AAV diagnosis was associated with increased risk of BSI at all times, with highest risk initially after first diagnosis. Interventions requiring a central venous catheter were solitary risk factors of BSI.
Read full abstract