Cardiothoracic & vascular anesthesia
Machine Learning–Based Prediction of Postoperative Vasoplegia After Cardiac Surgery
Authors' conclusion
A machine-learning model using routinely available perioperative and early postoperative data accurately predicted postoperative vasoplegia after cardiac surgery.
Study context
- Population
- Adults admitted to ICU after cardiac surgery, tertiary academic intensive care units (ICUs)
- Outcome
- Incident postoperative vasoplegia (CI >2.2 L/min/m², SVR <800 dyne·s/cm⁵ and MAP <65 mmHg, or CI >2.2 L/min/m² with norepinephrine plus epinephrine >0.2 µg/kg/min)
- Follow-up
- >6 to ≤48 hours after ICU arrival
Methodological detailsmodel development and internal validation, machine learning
- Model development and internal validation
- Machine learning
Main finding
As reported in the abstract
Prediction of postoperative vasoplegia, held-out test set
AUC of best-performing model (moderate discrimination)
0.74-0.75
Observational data: associations, not effects.
Further findings
Influential features
- Underlying hypertension, congestive HF, diabetes, male sex, preoperative angiotensin-converting enzyme inhibitor exposure
Original publication
Machine Learning–Based Prediction of Postoperative Vasoplegia After Cardiac Surgery. J Cardiothorac Vasc Anesth. Published 28 November 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →