Predictive Value of Serum Albumin Levels for Prognosis of Severe Sepsis in Late Preterm Neonates Admitted in NICU in CMH Abbottabad

Abstract

Background: Severe sepsis remains a critical cause of neonatal morbidity and mortality, particularly in late preterm infants. Identifying cost-effective prognostic markers is essential for timely risk stratification and intervention. While serum albumin has been linked to adverse outcomes, its role in predicting mortality among late preterm neonates with sepsis requires further validation. Objective: To determine the diagnostic accuracy of serum albumin levels for poor prognosis of severe sepsis in late preterm neonates taking mortality as gold standard. Study Design: Cross-sectional validation study. Duration and Place of Study: The study was carried out from January 2024 to September 2024 in the Neonatal Intensive Care Unit of Combined Military Hospital, Abbottabad. Methodology: A total of 311 late preterm neonates (34 0/7 to 36 6/7 weeks gestation) diagnosed with severe sepsis were included through non-probability consecutive sampling. Serum albumin levels were measured within 24 hours of NICU admission. Neonates were followed until discharge or death. Diagnostic metrics and ROC curve were calculated to evaluate the poor prognostic value of serum albumin. Results: The mean GA at birth and birth weight were 35.37 ± 0.85 weeks and 2101.55 ± 272.22 g, respectively. The average serum albumin level was 3.15 ± 0.43 g/dL. Mortality occurred in 71 neonates (22.8%). Serum albumin demonstrated a sensitivity of 87.32%, specificity of 78.33%, diagnostic accuracy of 80.39%, PPV of 54.39%, and NPV of 95.43%. Conclusion: Serum albumin is a valuable, accessible biomarker with strong negative predictive ability for mortality in late preterm neonates with sepsis.