A Risk Score for Predicting the Incidence of Hemorrhage in Critically Ill Neonates: Development and Validation Study

Area of Science:

• Neonatal critical care
• Hematology
• Medical prediction modeling

Background:

• Hemorrhage is a significant risk for critically ill neonates.
• Accurate prediction of bleeding events is crucial for timely intervention.
• Existing prediction tools may not fully integrate advanced hemostatic assessments.

Purpose of the Study:

• To develop and validate a prediction model for hemorrhage in critically ill neonates.
• To combine rotational thromboelastometry (ROTEM) parameters with clinical variables for enhanced prediction accuracy.
• To create an easily applicable risk score for quantifying 24-hour bleeding risk.

Main Methods:

• A cohort study of 332 critically ill neonates (full-term and preterm).
• Rotational thromboelastometry (ROTEM) testing and the neonatal bleeding assessment tool (NeoBAT) were utilized.
• Double selection least absolute shrinkage and selection operator (LASSO) logit regression was employed for model development, with cross-validation and bootstrap for internal validation.

Main Results:

• The Neonatal Bleeding Risk (NeoBRis) index was developed, incorporating EXTEM A10, LI60, platelet counts, and creatinine levels.
• The NeoBRis index demonstrated excellent predictive performance with an Area Under the Curve (AUC) of 0.908.
• Internal validation using bootstrap resampling confirmed the model's robustness with an AUC of 0.907.

Conclusions:

• An easy-to-apply prediction model (NeoBRis) for hemorrhage in critically ill neonates has been developed and internally validated.
• The model effectively combines ROTEM parameters and clinical data to predict 24-hour bleeding risk.
• Further external validation is needed, after which the NeoBRis index can aid clinicians in quantifying bleeding risk.