Triglyceride–glucose markers and obesity composites in relation to stroke diagnosis and mortality: A population-representative analysis

Lipid–glucose burden combined with adiposity distribution may influence stroke risk. Using nationally representative National health and nutrition examination survey (NHANES) 2005–2014 data, we examined associations between triglyceride-glucose (TyG)-based metabolic phenotypes and both stroke diagnosis and stroke mortality. The exposures included the TyG index and adiposity composites (TyG×BMI, TyG×WC), with TG/high-density lipoprotein cholesterol (HDL-C) and atherogenic index of plasma (AIP) as secondary exposures. Outcomes were baseline physician-diagnosed stroke and stroke mortality via National death index. Using complex survey weights, adjusted logistic and Cox models estimated per-SD odds/hazard ratios and quintile contrasts. Sex-stratified and spline analyses were conducted. The analytic sample approximately 11,500 participants, with 422 stroke diagnoses and 95 stroke deaths over a median follow-up of about 9.2 years. In Model 2, TyG×BMI and TyG×WC showed consistent positive associations with stroke diagnosis: per-SD ORs were 1.23 (95% CI 1.08–1.41; P=0.001) and 1.26 (95% CI 1.09–1.45; P=0.002), respectively. Associations for the standalone TyG index (OR 1.06; 0.93–1.20), TG/HDL-C (OR 1.10; 0.98–1.24), and AIP (OR 1.07; 0.96–1.20) were weaker or non-significant. Quintile and spline analyses supported approximately linear, monotonic increases from mid-to-high categories for TyG×BMI and TyG×WC. In sex-stratified analyses, associations were stronger and more consistent among women, while men’s associations were mainly driven by TyG×WC. No statistically significant associations were observed for stroke mortality in overall or stratified models. Using NHANES 2005–2014 linked to the 2019 NDI, we present the first sex-specific, integrated assessment of TyG and composite indices for stroke diagnosis and mortality, highlighting TyG×WC as a sensitive, low-cost marker that can enhance population screening and risk stratification.