Predicting short-term thromboembolic risk following Roux-en-Y gastric bypass using supervised machine learning
Hassam Ali1, Faisal Inayat2, Vishali Moond3
1Department of Gastroenterology, East Carolina University Brody School of Medicine, Greenville, NC 27834, United States.
World Journal of Gastrointestinal Surgery
|May 1, 2024
Summary
A new risk model identifies six key predictors for venous thromboembolism (VTE) after Roux-en-Y gastric bypass (RYGB). This tool aids clinicians in assessing VTE risk for RYGB patients, improving preoperative care.
Area of Science:
- Bariatric Surgery
- Cardiovascular Health
- Medical Informatics
Background:
- Roux-en-Y gastric bypass (RYGB) is effective for class III obesity but carries risks.
- Venous thromboembolism (VTE) is a significant cause of mortality post-RYGB.
- A need exists for a VTE risk assessment tool specific to RYGB patients.
Purpose of the Study:
- To develop and validate a scoring system for individualized 30-day VTE risk in RYGB patients.
- To identify key predictors of VTE following RYGB surgery.
Main Methods:
- Analysis of 6526 RYGB patients from the 2016-2021 MBSAQIP database.
- Backward elimination multivariate analysis to identify VTE predictors.
- Model validation using 5-fold cross-validation and receiver operating curves.
Main Results:
- Six predictors identified: COPD history, length of stay, prior DVT, HbA1c > 7%, venous stasis history, and preoperative anticoagulation.
- The risk model demonstrated good discriminatory power with an AUC of 0.79.
- The model showed satisfactory performance in sleeve gastrectomy and ESG procedures.
Conclusions:
- A simple, six-variable risk model can aid in preoperative VTE risk stratification for RYGB patients.
- The model provides insights into factors increasing VTE risk.
- This tool supports informed clinical decision-making for bariatric surgery candidates.
