Cross-population validation of the TyG-ABSI index as a novel predictor for chronic obstructive pulmonary disease: an integrated analysis using logistic regression and explainable machine learning
A simple combined metabolic score reliably predicts chronic lung disease risk across different populations and could guide preventive screening.
Cross-population validation in two independent cohorts (CHARLS China n=2,771 and NHANES USA n=1,925) demonstrates TyG-ABSI as a robust, independent COPD risk predictor with substantial gradient (OR 1.6-4.0 across quartiles), with tree-based ML models achieving AUC >0.96. The composite metabolic-adiposity index could serve as a simple screening tool for COPD risk in cardiometabolic patient populations.
What the study was
- Study design
- Cross-population validation; CHARLS (n=2,771) and NHANES (n=1,925 pre-diabetic participants); logistic regression + XGBoost/LightGBM/CatBoost/RF/KNN ML comparison; SHAP explainability
- Population
- Chinese community adults (CHARLS) and pre-diabetic US adults (NHANES)
- Sample size
- 4696
- Category
- Diagnostics
- Maturity
- Exploratory
- Journal
- BMC Pulmonary Medicine
Why it surfaced
Cross-population validation (Chinese + US cohorts) of a novel composite cardiometabolic-COPD risk index. Score 6 reflects cross-sectional design for NHANES component, unknown temporal precedence. AUC >0.96 should be viewed cautiously as likely optimistic for a prevalence-based model.
A plain-language summary of published research — not medical advice. Talk to a clinician about your own care.