Advancing necrotizing enterocolitis prediction through iterative monitoring

Area of Science:

• Neonatology
• Pediatric Gastroenterology
• Clinical Prediction Modeling

Background:

• Necrotizing enterocolitis (NEC) is a severe intestinal disease in preterm infants with high morbidity and mortality.
• Accurate prediction of NEC is challenging due to its complex, multi-factorial pathogenesis.
• Early prediction is crucial for effective management of critically ill neonates.

Purpose of the Study:

• To develop and validate a prediction model for NEC in preterm infants.
• To improve the accuracy of NEC prediction using repeated measurement data.
• To identify independent risk factors for NEC development.

Main Methods:

• Retrospective collection of clinical data from 300 premature infants (150 NEC, 150 non-NEC).
• Analysis of laboratory indicators on day 1, day 7, and day of NEC onset.
• Development of a nomogram using logistic regression to predict NEC, validated with concordance index and calibration plots.

Main Results:

• Antenatal steroids, antenatal antibiotics, pre-NEC probiotics, day 7 anion gap, and day 7 mean corpuscular volume were identified as independent risk factors.
• The developed NEC nomogram demonstrated strong predictive accuracy (ROC AUC 0.835).
• High concordance indices (0.835 training, 0.848 validation) and consistent calibration plots confirmed model efficacy.

Conclusions:

• The risk estimation nomogram for NEC provides significant clinical value.
• It aids in early prediction, targeted prevention, and timely intervention strategies for NEC.
• This tool can improve outcomes for preterm infants at risk of NEC.