Background: Background: Sepsis remains a leading cause of morbidity and mortality in preterm neonates. Early and accurate diagnosis is crucial for timely intervention. Procalcitonin (PCT) and C-reactive protein (CRP) are commonly used biomarkers for diagnosing neonatal sepsis, but their comparative diagnostic value in preterm neonates is not fully established.
Methods: We conducted a systematic review and meta-analysis of studies published between 2000 and 2025 evaluating the diagnostic accuracy of PCT and CRP for sepsis in preterm neonates. Databases searched included PubMed, Web of Science, and Scopus. Studies were included if they reported sensitivity, specificity, or area under the curve (AUC) for PCT and/or CRP in preterm neonates (<37 weeks gestational age) with suspected or confirmed sepsis. Diagnostic odds ratios (DOR), pooled sensitivity, specificity, and AUC were calculated using random-effects models. Heterogeneity was assessed with I² statistics.
Results: Twelve studies involving 1,842 preterm neonates (678 with confirmed sepsis) met inclusion criteria. Pooled sensitivity and specificity across all studies were 0.81 (95% CI: 0.75–0.86) and 0.85 (95% CI: 0.80–0.89) for PCT, and 0.78 (95% CI: 0.72–0.83) and 0.82 (95% CI: 0.77–0.86) for CRP. Overall AUC values were 0.88 for PCT and 0.83 for CRP. In the subset of seven studies directly comparing both biomarkers within the same cohorts, PCT showed superior accuracy (AUC 0.86 vs. 0.80, p=0.04). Subgroup analysis revealed higher diagnostic accuracy of PCT in LOS (AUC 0.89 vs. 0.82 for CRP).
Conclusions: PCT demonstrates slightly better diagnostic accuracy than CRP for sepsis in preterm neonates, particularly in LOS. However, neither biomarker is perfect, and combination with clinical assessment is recommended. Further studies with standardized cut-offs are needed.