53. Incidence of Bloodstream Infections and Outcomes in Patients with Severe COVID-19 Pneumonia

Background: Coronavirus disease 19 (COVID-19) leading to acute respiratory distress syndrome is associated with need for intensive care (IC), mechanical ventilation (MV), and prolonged recovery These patients are thus predisposed to blood stream infections which can worsen outcomes This risk may be aggravated by adjunctive therapies Methods: We reviewed the medical records of all adults admitted to Stony Brook University Hospital, NY, from March 1 to April 15, 2020 with severe COVID- 19 pneumonia (requiring high-flow O2) Patients who received MV or died within 24h were excluded Patients were followed until death or hospital discharge We reviewed positive blood cultures (PBC) for pathogenic microorganisms, and calculated the incidence of bacteremia, rates of infective endocarditis (IE), and impact on mortality Microbes isolated only once and belonging to groups defined as commensal skin microbiota were labelled as contaminants We also examined the impact of adjunctive therapies with immunosuppressive potential (steroids and tocilizumab), on bacteremia Results: A total of 469 patients with severe COVID-19 pneumonia were included (Table 1) Of these, 199 (42 4%) required IC and 172 (36 7%) MV Median length of stay was 13 days (8-22) and 94 (20 0%) had PBC Of these, 43 were considered true pathogens (bacteremia), with predominance of E faecalis and S epidermidis, and 51 were considered contaminants (Table 2) The incidence of bacteremia (43/469, 9 2%) was 5 1 per 1000 patient-days (95%CI 3 8-6 4) An echocardiogram was performed in 21 patients, 1 had an aortic valve vegetation (IE) by methicillin sensitive S aureus Bacteremia rates were nonsignificantly higher with steroids (5 9 vs 3 7 per 1000 patient-days;P=0 057) Use of tocilizumab was not associated with bacteremia (5 8 vs 4 8 per 1000 patient-days;P=0 28) Mortality was nonsignificantly higher in patients with (15/43, 34 9%) vs without (108/426, 25 4%) bacteremia (P=0 20) Length of stay was the strongest predictor of bacteremia, with risk increasing by 7% (95%CI 6%-9%, P< 0 001) per additional day Conclusion: The incidence of bacteremia was relatively low and IE was uncommon in this study of severe COVID-19 patients Risk of bacteremia increased with longer hospital stay and with steroids use, but not with tocilizumab (Table Presented)