BACKGROUNDSepsis is a significant cause of neonatal mortality. Both clinical and laboratory markers are often non- specific, and a blood culture contaminant may confuse management of the infant. This study aims to use an existing, inexpensive test, the neutrophil-to-lymphocyte ratio (NLR), to predict culture positivity and thus optimise antibiotic management.METHODSData on infants born ≤32 weeks gestation from 2015 to 2020 who were treated for sepsis were analyzed. Episodes were categorised based on culture positivity: i) positive with known pathogens, ii) true positive with coagulase-negative staphylococci (CoNS), iii) false positive with CoNS contaminants, iv) positive with other contaminants, and v) negative culture. NLR was compared with late C-reactive protein (CRP) and immature-to-total neutrophil ratio (ITR). Receiver operating curve (ROC) analysis was performed in this study. A cut- off value of 1.2 was chosen for NLR to detect culture.RESULTSA total of 2024 infants were included in the study, with no significant differences in demographics between groups. Infants with positive blood cultures of known pathogens had higher median NLR of 1.44 compared with those with contaminants (0.75) or negative cultures (0.84). A NLR cut-off value of 1.2 gave a sensitivity of 65.7 % and specificity of 63.7 %.The area under the curve (AUC) for NLR, late CRP and ITR were 0.71, 0.64 and 0.57 respectively. Combining NLR with late CRP would increase the AUC to 0.78 (95 % C·I 0.75-0.82, p < 0.01), whereas the addition of ITR to NLR and late CRP would not improve this. In a secondary analysis, between true and contaminant CoNS positivity, NLR and ITR could not be used to differentiate between these groups (p < 0.05 and 0.33), but late CRP could (p < 0.01).CONCLUSIONSClinicians should consider the use of NLR as an adjunct biomarker to late CRP in detecting culture positive sepsis in premature neonates. The use of ITR does not improve diagnostic accuracy for sepsis.