Accuracy of procalcitonin for diagnosing peripheral blood culture contamination among patients with positive blood culture for potential contaminants.

Abstract

Blood culture contamination is still a frequently observed event and may lead to unnecessary antibiotic prescriptions and additional hazards and costs. However, in patients hospitalized in tertiary care, true bacteremias for pathogens that are classically considered as contaminants can be observed. We assessed the diagnostic accuracy of procalcitonin for differentiating blood culture contamination from bacteremia in patients with positive blood cultures for potential contaminants. We carried out a retrospective, cross-sectional, observational study on consecutive patients hospitalized between January 2016 and May 2019 at the University Hospital of Nancy and who had a positive peripheral blood culture for a pathogen classically considered as a potential contaminant. During the study period, 156 patients were screened, and 154 were retained in the analysis. Among the variables that were significantly associated with a diagnosis of blood culture contamination in univariate analyses, four were maintained in multivariate logistic regression analysis: a number of positive blood culture bottles ≤ 2 (OR 23.76; 95% CI 1.94-291.12; P = 0.01), procalcitonin < 0.1ng/mL (OR 14.88; 95% CI 1.62-136.47; P = 0.02), non-infection-related admission (OR 13.00; 95% CI 2.17-77.73; P = 0.005), and a percentage of positive blood culture bottles ≤ 25% (OR 12.15; 95% CI 2.02-73.15; P = 0.006). These data provide new evidence on the usefulness of plasma procalcitonin as a reliable diagnostic biomarker in the diagnostic algorithm of peripheral blood culture contamination among patients hospitalized in tertiary care. ClinicalTrials.gov #NCT04573894.