Risk of New Bloodstream Infections and Mortality Among People Who Inject Drugs With Infective Endocarditis

IMPORTANCE: People who inject drugs (PWID) who are being treated for infective endocarditis remain at risk of new bloodstream infections (BSIs) due to ongoing intravenous drug use (IVDU). OBJECTIVES: To characterize new BSIs in PWID receiving treatment for infective endocarditis, to determine the clinical factors associated with their development, and to determine whether new BSIs and treatment setting are associated with mortality. DESIGN, SETTING, AND PARTICIPANTS: This retrospective cohort study was performed at 3 tertiary care hospitals in London, Ontario, Canada, from April 1, 2007, to March 31, 2018. Participants included a consecutive sample of all PWID 18 years or older admitted with infective endocarditis. Data were analyzed from April 1, 2007, to June 29, 2018. MAIN OUTCOMES AND MEASURES: New BSIs and factors associated with their development, treatment setting of infective endocarditis episodes (ie, inpatient vs outpatient), and 90-day mortality. RESULTS: The analysis identified 420 unique episodes of infective endocarditis in 309 PWID (mean [SD] patient age, 35.7 [9.7] years; 213 episodes [50.7%] involving male patients), with 82 (19.5%) complicated by new BSIs. There were 138 independent new BSIs, of which 68 (49.3%) were polymicrobial and 266 were unique isolates. Aerobic gram-negative bacilli (143 of 266 [53.8%]) and Candida species (75 of 266 [28.2%]) were the most common microorganisms. Ongoing inpatient IVDU was documented by a physician in 194 infective endocarditis episodes (46.2%), and 127 of these (65.5%) were confirmed by urine toxicology results. Multivariable time-dependent Cox regression demonstrated that previous infective endocarditis (hazard ratio [HR], 1.89; 95% CI, 1.20-2.98), inpatient treatment (HR, 4.49; 95% CI, 2.30-8.76), and physician-documented inpatient IVDU (HR, 5.07; 95% CI, 2.68-9.60) were associated with a significantly higher rate of new BSIs, whereas inpatient addiction treatment was associated with a significantly lower rate (HR, 0.53; 95% CI, 0.32-0.88). New BSIs were not significantly associated with 90-day mortality (HR, 1.76; 95% CI, 0.78-4.02); significant factors associated with mortality included inpatient infective endocarditis treatment (HR, 3.39; 95% CI, 1.53-7.53), intensive care unit admission (HR, 9.51; 95% CI, 4.91-18.42), and methicillin-resistant Staphylococcus aureus infective endocarditis (HR, 1.77; 95% CI, 1.03-3.03), whereas right-sided infective endocarditis was associated with a significantly lower mortality rate (HR, 0.41; 95% CI, 0.25-0.67). CONCLUSIONS AND RELEVANCE: In this study, new BSIs were common in PWID receiving parenteral treatment for infective endocarditis. Discharging patients to outpatient treatment was not associated with an increase in new BSI incidence or mortality; carefully selected PWID may therefore be considered for such treatment.