Identification of predictors responsible for neonatal sepsis and development of a diagnostic model

Abstract

Background

Neonatal sepsis must be diagnosed and treated earliest to avoid potential consequences. we aimed to develop a diagnostic model by identifying the risk factors responsible for neonatal sepsis and comparing the diagnostic values of C-reactive protein (CRP), Procalcitonin (PCT), and platelets.

Methods

We conducted a single-centre case-control study for five years on 300 neonates admitted to the neonatal intensive care unit (NICU). The study included neonates diagnosed with sepsis as the “case” and those without sepsis as the “control” group. Data regarding clinical and demographic characteristics were collected retrospectively from medical records. The risk factors responsible for neonatal sepsis were identified using logistic regression, and the diagnostic performance of biomarkers CRP, PCT and platelets were compared by determining their sensitivity, specificity and the AUC of the ROC curve. Subsequently, the diagnostic model for neonatal sepsis was brought about using the identified risk factors and values of CRP, PCT and platelets.

Results

CRP was the biomarker studied with the highest sensitivity (65.32 %) and specificity (85.37 %) among CRP, PCT and platelets. The highest AUC value (0.811) on ROC curve analysis was also exhibited by CRP. Additionally, the independent risk factors identified for the development of neonatal sepsis were the bodyweight category, gestational age category, and CRP levels (p < 0.05) using multivariate logistic regression.

Conclusion

CRP is the most dependable predictor for the diagnosis of neonatal sepsis in the study. Factors like body weight and gestational age contribute to the development of neonatal sepsis.