CBC ratios for early sepsis risk stratification in newborns: EOS/LOS versus non-septic controls

Despite advances in neonatal intensive care units (NICUs), neonatal sepsis remains a leading cause of morbidity and mortality worldwide. Early and accurate diagnosis is challenging because clinical signs are often nonspecific and no single biomarker has shown perfect sensitivity and specificity. Recently, complete blood count–based parameters such as the neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) have gained attention as simple and low-cost markers of systemic inflammatory response. This study was designed to assess the diagnostic utility of NLR and PLR in early-onset sepsis (EOS) and late-onset sepsis (LOS), in comparison with jaundiced and healthy neonates, and to evaluate their performance relative to procalcitonin (PCT) and C-reactive protein (CRP). In this study, we reviewed the records of neonates hospitalized in a level III NICU between March 2022 and March 2025. A total of 446 infants were classified into four groups: EOS (first 3 postnatal days), LOS (≥ 4th day), jaundice without sepsis, and healthy controls. Pre-treatment laboratory data, including complete blood count, CRP, and PCT, were extracted. The neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) were calculated as the ratios of absolute neutrophil and platelet counts to lymphocyte counts, respectively, and group differences were analyzed using nonparametric tests. Receiver operating characteristic (ROC) curve analysis was used to assess the diagnostic performance of each biomarker for differentiating sepsis-positive (EOS + LOS) from sepsis-negative (jaundice+healthy) infants. Of the 446 neonates, 143 (32.1%) had EOS, 101 (22.6%) had LOS, 102 (22.9%) were in the jaundice group, and 100 (22.4%) were healthy controls. Median NLR (IQR) differed across groups and was highest in EOS (1.97 [1.10–2.78]) compared with LOS (0.71 [0.46–1.43]), jaundice (0.84 [0.59–1.56]), and healthy controls (0.67 [0.46–1.19]) (p < 0.001). There was no significant variation in PLR among the four groups (p = 0.133). In ROC analysis for discriminating sepsis-positive (EOS + LOS) from sepsis-negative (jaundice+healthy) infants, PCT showed the highest diagnostic accuracy (AUC 0.970), followed by CRP (AUC 0.747) and NLR (AUC 0.650), whereas PLR had limited discriminative ability (AUC 0.543). NLR may serve as a supportive, low-cost marker for neonatal sepsis—particularly in EOS—when interpreted alongside PCT, CRP, and clinical findings. In this cohort, PLR did not provide meaningful additional diagnostic value. Potential implications for antibiotic stewardship should be confirmed in prospective studies that directly assess prescribing and treatment duration.