<![CDATA[Background: Early-onset neonatal sepsis remains a critical cause of mortality in developing nations. Blood culture, the gold standard, suffers from delay and low sensitivity. While hematologic indices such as neutrophil-to-lymphocyte ratio (NLR) are used in adults, their utility in the first 72 hours of life is confounded by physiological instability. This study evaluates the diagnostic accuracy of the platelet-to-lymphocyte ratio (PLR) compared to NLR, mean platelet volume (MPV), and red cell distribution width (RDW) in early-onset neonatal sepsis. Methods: A retrospective observational study was conducted on 55 neonates (25 septic, 30 symptomatic non-septic controls) at a tertiary center in Indonesia. Sepsis was defined by clinical criteria and C-reactive protein positivity, independent of complete blood count parameters, to avoid incorporation bias. Diagnostic performance was assessed using Mann-Whitney U tests, receiver operating characteristic curve analysis, and multivariable logistic regression to control for confounders, including asphyxia. Results: The median PLR was significantly lower in the sepsis group compared to controls (32.6 [IQR 3.4–100.4] versus 71.1 [IQR 45.3–82.9]; p = 0.016). Conversely, NLR (p = 0.80), MPV (p = 0.163), and RDW (p = 0.422) showed no significant discrimination. PLR yielded an area under the curve of 0.724. At a cut-off of equal to or less than 40.5, determined by the Youden Index, PLR demonstrated a sensitivity of 68.0%, specificity of 73.3%, positive likelihood ratio of 2.55, and negative likelihood ratio of 0.44. Multivariable regression confirmed PLR as an independent predictor (Adjusted Odds Ratio 0.96; 95% CI 0.93–0.99; p = 0.038) after adjusting for birth asphyxia. Conclusion: PLR demonstrates superior discriminative ability over NLR for early-onset sepsis in this cohort. The distinct inverse PLR phenomenon reflects sepsis-induced thrombocytopenia and bone marrow suppression. While not a standalone diagnostic tool, PLR serves as a valuable, zero-cost adjunctive marker for risk stratification in resource-limited settings.]]>
Copyrights © 2026
