Pediatric cardiac surgeries involving cardiopulmonary bypass (CPB) often necessitate perioperative red blood cell transfusions (PRT) due to increased surgical complexity, blood loss, and metabolic demands. Despite its benefits, PRT carries risks such as prolonged hospital stays and postoperative complications. Predicting the need for transfusion in pediatric patients remains a challenge, necessitating a robust clinical tool to improve decision-making and optimize blood resource utilization.
This study aimed to develop and validate a nomogram-based predictive model for PRT risk in children undergoing cardiac surgery with CPB. Researchers analyzed data from 19,155 pediatric patients treated between 2014 and 2021 at Fuwai Hospital in China. The dataset was randomly divided into a training cohort (70%) and a testing cohort (30%). Predictor selection was conducted using univariate logistic regression, least absolute shrinkage and selection operator (LASSO) regression, and multivariate logistic regression.
Eight key predictors of PRT were identified:
- Age
- Weight
- Preoperative hemoglobin levels
- Presence of cyanotic congenital heart disease (CCHD)
- CPB duration
- Minimum rectal temperature during CPB
- CPB priming volume
- Use of a small incision
The final predictive model demonstrated strong discrimination, with an area under the receiver operating characteristic curve (AUC) of 0.886 in the training cohort and 0.883 in the testing cohort. Calibration plots indicated a close alignment between predicted and actual outcomes, and decision curve analysis confirmed its clinical benefit.
To enhance practical use, the model stratified patients into three risk categories:
- Low-risk: <30% probability of transfusion
- Intermediate-risk: 30–70% probability
- High-risk: >70% probability
Patients in the high-risk category had a 91.6% chance of requiring PRT, highlighting the model’s ability to effectively differentiate transfusion needs.
The study underscores the importance of integrating predictive modeling into pediatric cardiac surgery planning. Clinicians can use the nomogram to implement targeted blood conservation strategies, such as intraoperative cell salvage and tailored priming solutions, thereby minimizing unnecessary transfusions. The model provides a valuable tool for optimizing perioperative blood management, improving patient outcomes, and reducing healthcare costs.
Despite its strengths, the study has limitations. It was based on a single-center retrospective dataset, requiring external validation for broader applicability. Additionally, some intraoperative variables such as CPB duration—may limit its preoperative utility. Future research should focus on refining the model with additional clinical parameters and expanding validation efforts.
