698 - Utility of multiple blood cultures in the detection of neonatal sepsis

Sepsis continues to be a leading cause of infant mortality and morbidity. Positive blood cultures are the gold standard detection of infection, however there is no consensus on the amount and number of blood cultures for evaluation of sepsis ( 2 vs 1 blood culture). The current practice at Level 3 NICU Parkland Hospital- Dallas, is to draw 2 samples of 2 different sites for sepsis evaluation. Prior to 2019, volume of 0.5 ml blood culture was sent however current practice since 2019 has been to draw 1 ml blood cultures to improve sensitivity of pathogen detection.


Objective:

To evaluate the utility of 1 versus 2 blood cultures in the detection of true infection in neonates.


Design/Methods: Each sepsis evaluation was considered as an event. True sepsis was considered based on the organism detected and duration of antibiotics. Positive blood culture was considered as contaminant if the organism detected was not pathogenic. A retrospective chart review was conducted from 2014-2019, for all sepsis evaluation at the NICU Parkland hospital- Dallas. Data from 2020-2021 was evaluated to if percentage of contamination decreased with increasing blood culture volumes from 0.5 ml to 1 ml.

Results:

A total of 7456 events were identified from year 2014-2019 with 285 positive blood culture events (table 2.1). Events with < 2 or >2 blood culture samples were excluded. 97 events were classified as EOS and 188 events as LOS. In EOS events with 1 positive blood culture, 40 events were treated as true sepsis and 38 events as contaminants. 19 EOS events with 2+ positive blood culture was treated as true sepsis.  In LOS, events with 2 positive blood cultures, 40 were treated as sepsis and 20 events as contaminants.74 LOS events were treated as true sepsis with 1 positive blood culture and 54 events as contaminants. There is around 50% chance of missing infection in EOS and ~ 45% in LOS if only 1 blood c/s sample is sent. Providers discontinued antibiotics ~50% in EOS and ~ 40% in LOS with 1+blood culture.67% LOS events with 1 positive blood culture with CONS were treated as infection which may have been a contaminant. Increasing the volume of blood culture did not decrease the rate of contamination in EOS events and no change noted in LOS events (Table 2.2)


Conclusion(s): The practice of obtaining 2 blood cultures, irrespective of blood volumes during a septic evaluation increases the sensitivity to detect true sepsis and reduces the use of antibiotic usage by ruling out infection if culture is considered a contaminant