Abstract
Background/Objectives: Neonatal sepsis remains a major diagnostic challenge, particularly in postoperative infants where systemic inflammatory responses may mimic infection; therefore, reliable biomarkers are urgently needed to distinguish sepsis from normal post-surgical changes. Methods: This prospective cohort study included 135 neonates who underwent surgery without preoperative signs of infection. Serum levels of C-reactive protein (CRP) and procalcitonin (PCT) were measured preoperatively and at 24, 72 and 120 h postoperatively. Patients were classified as having proven sepsis based on positive blood cultures or clinical sepsis using European Medicines Agency (EMA) neonatal sepsis scoring. Biomarker levels were compared between septic and non-septic infants, and diagnostic performance was evaluated using receiver operating characteristic analysis. Results: Sixteen infants had proven sepsis and twenty-five had clinical sepsis. Both CRP and PCT levels showed a significant rise in septic infants at 72 h postoperatively (p < 0.01) and remained elevated at 120 h. PCT demonstrated the highest diagnostic accuracy at 72 h (AUC 0.911; threshold 1.2 µg/L; sensitivity 83%; specificity 84%), while CRP also showed good performance (AUC 0.802-0.838). Conclusions: CRP, PCT and albumin are valuable postoperative sepsis markers; however, single-parameter interpretation is insufficient, and a multiparametric, time-dependent approach provides superior diagnostic reliability in neonatal sepsis.