227. Epidemiology of bloodstream infections in Veterans Health Administration, 2010-2020

Abstract Background National trends of bloodstream infections (BSI), their etiologies, and prevalence of resistance are not well described. We reviewed BSI during 2010-2020 in the Veterans Health Administration (VHA), the largest healthcare system in the United States. Methods Demographic, microbiological, and healthcare exposure data were extracted from VHA databases. A case was defined as isolation of a microbe from blood specimens collected from a hospitalized person; common commensals required matching organisms isolated within two consecutive days. The first organism-specific episode within a 14-day period was counted. Staphylococcus, Enterococcus, S. pneumoniae, and gram-negative isolates were assessed for resistance to methicillin, vancomycin, any antimicrobial, and extended-spectrum cephalosporins or carbapenems, respectively. Cases were classified as community acquired (CA-), healthcare-associated community onset (HCO-), and hospital onset (HO-). Trends were estimated by generalized linear mixed models. Results During 2010-2020, incidence of CA-BSI decreased from 42.2 to 27.6 per 100,000 users, HCO-BSI decreased from 63.7 to 40.7 per 100,000 users, and HO-BSI decreased from 28.2 to 16.4 per 100,000 users (Figure 1A). S. aureus and E. coli were the most common in CA-BSI and HCO-BSI; S. aureus and Enterococcus were the most common in HO-BSI; the prevalence of E. coli increased in BSI across classifications (Figure 1B). Incidence of BSI caused by resistant Pseudomonadales and Enterococcus decreased by more than 15% annually; annual incidence of BSI caused by other organisms decreased by less than 10% or remained unchanged with the exception of extended-spectrum cephalosporin resistant E. coli, which increased 6% annually (Figure 2). HO-BSI were more resistant than CA-BSI and HCO-BSI across organisms; resistance among E. coli and S. pneumoniae BSI increased across classifications (Figure 3). Figure 1. Trends of bloodstream infections by organism in Veterans Health Administration, 2010-2020. (A) Incidence per 100,000 users. (B) Percentage of incident BSI by organism. Trends are adjusted for distributions of age, gender, and number of users, in addition to accounting for clustering by county and facility. Community acquired: positive culture collected less than 4 days after hospitalization from a person without previous healthcare exposures. Healthcare-associated community onset: positive culture collected less than 4 days after hospitalization from a person with previous healthcare exposures. Hospital onset: positive culture collected 4 or more days after hospitalization. Figure 2. Percentage change in annual incidence of bloodstream infections by organism in Veterans Health Administration, 2010-2020. Dots represent point estimates and horizontal bars represent 95% confidence intervals. Figure 3. Trends in prevalence of resistance among organisms causing bloodstream infection by epidemiological classification in Veterans Health Administration, 2010-2020 Trends are adjusted for distributions of age, gender, and number of users, in addition to accounting for clustering by county and facility. Community acquired: positive culture collected less than 4 days after hospitalization from a person without previous healthcare exposures. Healthcare-associated community onset: positive culture collected less than 4 days after hospitalization from a person with previous healthcare exposures. Hospital onset: positive culture collected 4 or more days after hospitalization. Conclusion BSI incidence decreased during 2010-2020 across classifications. CO-BSI and HCO-BSI occurred more frequently and were less resistant than HO-BSI. S. pneumoniae and E. coli BSIs became more resistant over time. Increasing incidence of BSI caused by E. coli resistant to extended-spectrum cephalosporins warrants urgent investigation. Disclosures All Authors: No reported disclosures