Published online by Cambridge University Press: 07 November 2019
Healthcare-associated bloodstream infections (HABSIs) are a significant cause of mortality and morbidity in the neonatal intensive care unit (NICU) population. Our objectives were to review the epidemiology of HABSIs in our NICU and to examine the applicability of National Healthcare Safety Network (NHSN) definitions to the NICU population.
We performed a retrospective review of all neonates admitted to the 54-bed, level IV NICU at Yale-New Haven Children’s Hospital with a HABSI between January 1, 2013, and December 31, 2018. Clinical definitions per NICU team and NHSN site-specific definitions used for source identification were compared using the McNemar χ2 test.
We identified 86 HABSIs with an incidence rate of 0.80 per 1,000 patient days. Only 13% of these were CLABSIs. Both CLABSIs and non–catheter-related bloodstream infections occurred primarily in preterm neonates, but the latter were associated with a significantly higher incidence of comorbidities and the need for respiratory support. The NHSN definitions were less likely to identify a source compared to the clinical definitions agreed upon by our NICU treating team (P < .001). Furthermore, 50% of patients without an identified source of infection by NHSN definitions were bacteremic with a mucosal barrier injury organism, likely from gut translocation.
HABSIs occur primarily in premature infants with comorbidities, and CLABSIs account for a small proportion of these infections. With the increasing focus on HABSI prevention, there is a need for better NHSN site-specific definitions for the NICU population to prevent misclassification and direct prevention efforts.
PREVIOUS PRESENTATION: These data were presented as poster #585, “Shifting Focus Towards Healthcare-Associated Bloodstream Infections: The Need for More NICU-Specific NHSN Definitions,” at IDWeek 2019 on October 3, 2019, in Washington, DC.
Full text views reflects PDF downloads, PDFs sent to Google Drive, Dropbox and Kindle and HTML full text views.