Clinical consequences of contaminated blood cultures in adult hospitalized patients at an institution utilizing a rapid blood-culture identification system

Abstract

To assess the clinical impact of contaminated blood cultures in hospitalized patients during a period when rapid diagnostic testing using a FilmArray Blood Culture Identification (BCID) panel was in use. Retrospective cohort study. Single academic medical center. Patients who had blood culture(s) performed during an admission between June 2014 and December 2016. Length of hospital stay and days of antibiotic therapy were assessed in relation to blood-culture contamination using generalized linear models with univariable and multivariable analyses. Among 11,474 patients who had blood cultures performed, the adjusted mean length of hospital stay for patients with contaminated blood-culture episodes (N = 464) was 12.3 days (95% confidence interval [CI], 11.4-13.2) compared to 11.5 days (95% CI, 11.0-11.9) for patients (N = 11,010) with negative blood-culture episodes (P = .032). The adjusted mean durations of antibiotic therapy for patients with contaminated and negative blood-culture episodes were 6.0 days (95% CI, 5.3-6.7) and 5.2 days (95% CI, 4.9-5.4), respectively (P = .011). Despite the use of molecular-based, rapid blood-culture identification, contamination of blood cultures continues to result in prolonged hospital stay and unnecessary antibiotic therapy in hospitalized patients.