Work

Association between triglyceride glucose-body mass index and MASH, cardiovascular disease in MASLD patients: a cross-sectional study and machine learning analysis

Year: 2025

Type: article

Abstract: Background: The triglyceride glucose-body mass index (TyG-BMI) is a useful marker for metabolic dysfunction-associated steatotic liver disease (MASLD) and cardiovascular disease (CVD). However, its ability to differentiate metabolic dysfunction-associated steatohepatitis (MASH) and CVD risk among patients with MASLD requires further investigation. This study evaluates the association between TyG-BMI and MASH/CVD risk in patients with MASLD and applies machine learning (ML) to identify relevant risk factors. Methods: The study used data from National Health and Nutrition Examination Survey (NHANES) to analyzed associations between TyG-BMI and two clinical outcomes: MASH, defined by the FibroScan-AST (FAST) score ≥0.35, and self-reported CVD. Participants were categorized based on TyG-BMI quartiles: Q1 (<246.79), Q2 (246.79-280.32), Q3 (280.33-323.99), and Q4 (≥324.00). Multivariate survey-weighted logistic regression and restricted cubic splines (RCS) were used to assess relationships and potential nonlinearity. Multiple ML models were employed, with feature importance interpreted via Shapley Additive Explanations (SHAP) analysis. Results: Among 674 MASLD participants (390 males), higher TyG-BMI was independently associated with increased risks of MASH and CVD risk. Compared with Q1, Q4 had adjusted odds ratios of 24.46 (95 % CI: 2.94-203.31, P = 0.003) for MASH and 3.53 (95 % CI: 1.26-9.90, P = 0.017) for CVD. RCS indicated linear relationships between TyG-BMI and both outcomes. The gradient boosting machine and support vector machine exhibited the optimal performance best in discriminating high-risk MASH (ROC: 0.910) and CVD (ROC: 0.773), confirming TyG-BMI as a significant risk factors. Conclusion: TyG-BMI effectively identifies MASH and CVD risk in patients with MASLD, offering clinicians a practical tool for risk stratification and management.

Source: Metabolism Open

DOI: 10.1016/j.metop.2025.100423