Interpretable machine learning and incremental prognostic model for all-cause and cardiovascular mortality in obesity with MASLD population: Comorbidity insights from triglyceride-glucose and its obesity-related derivatives
Randomized trial investigates mortality associations in obesity with MASLD, highlighting predictive value of TyG.
Key Points
This research aims to explore the prognostic value of triglyceride-glucose and its derivatives for mortality in obesity associated with metabolic dysfunction.
Analyzed data from 4807 adults with obesity and MASLD from National Health and Nutrition Examination Survey (1999-2018).
Utilized Kaplan-Meier, Cox regression, and restricted cubic splines to assess mortality associations.
Developed predictive models using machine learning and SHAP for feature importance interpretation.
TyG and TyG-WHtR were associated with all-cause mortality; TyG-WHtR showed a 31.7% increase in risk per unit increase.
For cardiovascular mortality, TyG-WHtR demonstrated a 46.7% risk increase per unit increase.
XGBoost identified age, fasting plasma glucose, and male gender as key risk factors, enhancing prediction performance when integrating TyG.