641. Clinical Predictors of Hospital-Acquired Bloodstream Infections

Abstract Background Hospital-acquired bloodstream infections (HABSI) are associated with increased mortality and decreased hospital quality metrics. This has led to an increased focus on blood culture stewardship. Little data exists regarding predictive factors of bacteremia in hospitalized patients. We aim to determine what clinical characteristics in patients were predictive of HABSI. Methods This is a retrospective case-control study of 540 patients with positive blood cultures admitted to our health system between September 1, 2017, to April 1, 2020. Electronic medical records of patients with positive blood cultures were independently reviewed to determine contamination versus true bacteremia. We looked at different clinical parameters and laboratory investigations within 24 hours of drawing blood cultures. Clinical variables were age ≥ 60 years, heart rate ≥ 90/minute, systolic blood pressure ≤ 90 mmHg or use of a vasopressor, oral temperature > 38°Celsius (100.4°Fahrenheit), white blood cells (WBC) count ≥12,000/µL, lymphocytes ≤ 1000/mm3, platelets < 150,000 /µL, and creatinine >2.0 mg/dL. Stepwise logistic regression analysis was used for predictive statistical model development. Results In a cohort of 481 patients with hospital-acquired bacteremia, 350 cases had true bacteremia and 131 cases were contaminated blood cultures. Stepwise regression analysis showed that white blood cell (WBC) count ≥ 12,000 cells/µL, lymphocyte count ≤ 1000/mm3, creatinine > 2.0 mg/dL, and oral temperature > 38°C (100.4°F) were associated with HABSI (R-square= 0.06, p value= 0.002). Conclusion Our findings suggest that WBC count, lymphocyte count, creatinine, and oral temperature together can be used to develop appropriate blood culture stewardship models in the inpatient setting. This may help minimize unnecessary blood cultures. Disclosures All Authors: No reported disclosures


Clinical Predictors of Hospital-Acquired Bloodstream Infections
Radhika Sheth, MD 1 ; Mehakmeet Bhatia, DO 1 ; Vivek Kak, MD 1 ; 1 Henry Ford Allegiance Health, Jackson, Michigan Session: P-29. Diagnostics: Bacteriology/mycobacteriology Background. Hospital-acquired bloodstream infections (HABSI) are associated with increased mortality and decreased hospital quality metrics. This has led to an increased focus on blood culture stewardship. Little data exists regarding predictive factors of bacteremia in hospitalized patients. We aim to determine what clinical characteristics in patients were predictive of HABSI.
Methods. This is a retrospective case-control study of 540 patients with positive blood cultures admitted to our health system between September 1, 2017, to April 1, 2020. Electronic medical records of patients with positive blood cultures were independently reviewed to determine contamination versus true bacteremia. We looked at different clinical parameters and laboratory investigations within 24 hours of drawing blood cultures. Clinical variables were age ≥ 60 years, heart rate ≥ 90/minute, systolic blood pressure ≤ 90 mmHg or use of a vasopressor, oral temperature > 38°Celsius (100.4°Fahrenheit), white blood cells (WBC) count ≥12,000/ µL, lymphocytes ≤ 1000/mm3, platelets < 150,000 /µL, and creatinine >2.0 mg/ dL. Stepwise logistic regression analysis was used for predictive statistical model development.
Results. In a cohort of 481 patients with hospital-acquired bacteremia, 350 cases had true bacteremia and 131 cases were contaminated blood cultures.
Conclusion. Our findings suggest that WBC count, lymphocyte count, creatinine, and oral temperature together can be used to develop appropriate blood culture stewardship models in the inpatient setting. This may help minimize unnecessary blood cultures.
Disclosures. Background. Carbapenem (Carb) minimum inhibitory concentration (MIC) breakpoints were lowered by CLSI in 2010 and recognized by FDA in 2012. Adoption of revised breakpoints is often slow, which may lead to under-reporting of Carb non-susceptibility (NS) by facilities. We compare facility-reported rates of Carb-NS ENT to the CLSI MIC breakpoints for a large nationwide collection of isolates in the United States (US) from 2016-2019.
Results. Overall,77.4% (937,926/1,211,845) and 90.6% (2,157,785/2,381,824) non-duplicate ENT isolates with facility-reported susceptibility results also had interpretable MIC results for ETP and IPM/MEM/DOR, respectively (Tables). ETP S rates were 99.3% and 99.1% as reported by facilities and using CLSI criteria, respectively. S rates of other Carbs were 98.9% and 98.4% by facility reporting and CLSI criteria, respectively. Systematic application of CLSI breakpoints under-reported EPT-I and -R isolates by 24.2% and 16.4%, respectively, and identification of IPM/MEM/DOR-I and -R isolates by 31.3% and 22.7%, respectively. Background. Bloodstream infections are a major cause of morbidity and mortality in hospitalized patients. Prompt initiation of effective antimicrobials are essential to optimize patient outcomes. New diagnostic technologies rapidly identifying bacteria, viruses, fungi, and parasites in infections of various body sites. There is a paucity of literature determining if stewardship programs run by one trained pharmacist with rapid diagnostics decreases time to optimal antimicrobial therapy.
Methods. This was a retrospective chart review of positive bloodstream infections identified via rapid diagnostic technologies. The EHR of admitted adult patients