Related Experiment Video
Updated: Jun 2, 2026
04:00
Identifying Frailty Using Point-of-Care Ultrasonography: Image Acquisition and Assessment
Published on: July 26, 2024
Integrating sarcopenia into ICU-acquired weakness risk stratification: a machine learning-based prediction model for
Peng Zheng1, Xinwei Liu1, Chunxia Zhang1
1Department of Intensive Care Unit, Clinical Oncology School of Fujian Medical University, Fujian Cancer Hospital, NHC Key Laboratory of Cancer Metabolism, Fuzhou, Fujian, China.
Frontiers in Nutrition
|June 1, 2026
Summary
Sarcopenia predicts intensive care unit-acquired weakness (ICU-AW) in critically ill patients. An XGBoost model integrating sarcopenia improves early risk stratification for ICU-AW.
Area of Science:
- Critical Care Medicine
- Geriatrics
- Biostatistics
Background:
- Sarcopenia, muscle loss, is linked to weakness in elderly and chronically ill individuals.
- Its role in predicting ICU-acquired weakness (ICU-AW) among critically ill patients requires clarification.
Purpose of the Study:
- To determine if sarcopenia predicts ICU-AW in critically ill patients.
- To develop a machine learning model incorporating sarcopenia for ICU-AW risk stratification.
Main Methods:
- Retrospective analysis of 1,324 critically ill patients, split into training (n=927) and validation (n=397) sets.
- Sarcopenia assessed via L3 skeletal muscle area on CT scans; LASSO and Boruta algorithms identified predictors.
- Ten machine learning models were developed, with XGBoost selected for optimal performance and SHAP analysis for feature importance.
Main Results:
- Six key predictors identified: age, APACHE II score, sarcopenia, sepsis, mechanical ventilation, and lactic acid.
- The XGBoost model achieved high predictive performance (AUC 0.838 training, 0.843 validation).
- Sarcopenia was the third most important predictor in the XGBoost model, after APACHE II and age.
Conclusions:
- Sarcopenia is a significant predictor of ICU-AW in critically ill patients.
- Integrating sarcopenia into the XGBoost model enhances early risk stratification for ICU-AW.