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

Abstract

The aim of the study was to develop and validate a prediction model for hemorrhage in critically ill neonates which combines rotational thromboelastometry (ROTEM) parameters and clinical variables. This cohort study included 332 consecutive full-term and preterm critically ill neonates. We performed ROTEM and used the neonatal bleeding assessment tool (NeoBAT) to record bleeding events. We fitted double selection least absolute shrinkage and selection operator logit regression to build our prediction model. Bleeding within 24 hours of the ROTEM testing was the outcome variable, while patient characteristics, biochemical, hematological, and thromboelastometry parameters were the candidate predictors of bleeding. We used both cross-validation and bootstrap as internal validation techniques. Then, we built a prognostic index of bleeding by converting the coefficients from the final multivariable model of relevant prognostic variables into a risk score. A receiver operating characteristic analysis was used to calculate the area under curve (AUC) of our prediction index. EXTEM A10 and LI60, platelet counts, and creatinine levels were identified as the most robust predictors of bleeding and included them into a Neonatal Bleeding Risk (NeoBRis) index. The NeoBRis index demonstrated excellent model performance with an AUC of 0.908 (95% confidence interval [CI]: 0.870–0.946). Calibration plot displayed optimal calibration and discrimination of the index, while bootstrap resampling ensured internal validity by showing an AUC of 0.907 (95% CI: 0.868–0.947). We developed and internally validated an easy-to-apply prediction model of hemorrhage in critically ill neonates. After external validation, this model will enable clinicians to quantify the 24-hour bleeding risk.

Reply to Ghirardello et al Letter to the Editor

“…Regarding the comment that we "did not provide clinical data to better characterize the severity of illness for patients enrolled" in our work, the study design and the inclusion criteria have been clearly reported in the Methods section. 2 We will be happy to provide specific additional clinical data if available in our dataset.…”

“…We thank Dr. Ghirardello and colleagues 1 for their interest in our study, which introduced an easy-to-apply, rotational thromboelastometry-based, prediction model of hemorrhage in critically ill neonates. 2 Their letter gives us the opportunity to address their comments and provide clarifications to their assumptions.…”

Prospective Temporal Validation of the Neonatal Bleeding Risk (NeoBRis) Index

“…Prediction scores and bleeding assessment tools, which could help estimating the risk/benefit ratio of transfusions, are crucial for supporting clinical decisions and bleeding management in this vulnerable population [1][2][3][4]. Recently, our research group developed and published in the journal a multivariable prediction tool for 24hour bleeding risk in critically-ill neonates (Neonatal Bleeding Risk [NeoBRis] index) [5]. This assessment tool includes extrinsically activated rotational thromboelastometry parameters (EXTEM A10 and LI60 [amplitude recorded at 10 minutes and Lysis Index at 60 minutes, respectively]), platelet count and creatinine plasma levels and has demonstrated excellent performance (AUC: 0.908; 95% CI: 0.870-0.946).…”

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

“…If this is indeed true, the percentage of IVH exceeds 50% in this latter population, which is extraordinarily high for the population included in the most recent study. 1 Third, the authors did not provide clinical data (i.e., Apgar score, need for mechanical ventilation and inotropes, SNAPPE score, mortality) to better characterize the severity of illness for the patients enrolled that could justify such a high incidence of IVH observed in their cohort.…”

“…We read with interest the study by Sokou et al 1 and would like to make comments. The authors proposed a risk score for predicting hemorrhage of any grade in critically ill neonates within 24 hours of rotational thromboelastometry testing, using the Neonatal Bleeding Assessment Tool (NeoBAT) score.…”