Assistant Professor of Pediatrics Yale School of Medicine New Haven, Connecticut, United States
Background: Blood culture remains the gold standard for diagnosing bacteremia in neonates. The practice of obtaining blood cultures from two separate sites is commonly used in the neonatal intensive care unit (NICU) in the evaluation for late onset sepsis (LOS), to differentiate true infections from contaminants. However, no clear guidance exists regarding the utility of this practice in the evaluation for early-onset sepsis (EOS).
Objective: To determine the utility of obtaining two separate-site blood cultures in the diagnosis and treatment of EOS in the NICU.
Design/Methods: Since 2009, standard practice in our NICU has been to attempt two separate site blood cultures when evaluating for EOS. Clinical and microbiological data were collected retrospectively on all neonates hospitalized in the NICU at Yale New Haven Children’s Hospital (YNHCH) who underwent an EOS evaluation from 2013-2021. EOS was defined as having at least one positive blood culture, obtained at < 72 hours of age, and growing either a true neonatal pathogen or a commensal pathogen treated with antimicrobials for >5 days. When two blood cultures were obtained, discordant and concordant results were assessed and compared.
Results: 2298 EOS evaluations were performed in the YNHCH NICU among 43,078 live births (LB) from 2013-2021. Two blood cultures were obtained from separate sites in 66% of evaluations. Fifty-seven suspected cases of EOS were identified (1.3 suspected cases per 1000 LB). In 11 (19%), only a single culture was obtained, with all deemed as true infections. These included five cases attributed to Escherichia coli (E. coli), two to Group B streptococcus (GBS), and two to common contaminant species treated based on lack of a second culture and clinical concerns. Two separate site blood cultures were obtained in 46 (81%) positive cases, 25 (54%) of which yielded discordant results (e.g., one positive and one negative culture) with 12 (48%) assessed as true infections (six due to E. coli and three to GBS) and 13 (52%) as contaminants. The overall positivity rate comparing two versus single site blood cultures was 3.1% versus 1.4% (p=0.014). The true positive rate comparing two versus single site blood cultures was 2.1% versus 1.4% (p=0.216).Conclusion(s): A two-site blood culture practice for diagnosing EOS yielded a relatively high rate of discordance, with nearly half deemed as true cases of EOS. As with LOS, the use of two blood cultures may also assist in differentiating contamination from true infections, potentially reducing unnecessary antibiotic utilization.