Comparative associations of TyG, TyG-BMI, TyG-WC, and TyG-WWI with blood pressure phenotypes and arterial stiffness: findings from NHANES 1999–2018

Insulin resistance and adverse adiposity distribution are associated with elevated blood pressure. Although the triglyceride-glucose (TyG) index is widely used as a surrogate marker of insulin resistance, it does not directly capture body-shape-related characteristics. This study compared the associations and sample-based discrimination of the triglyceride-glucose-weight-adjusted waist index (TyG-WWI), the conventional TyG index, and other TyG-derived anthropometric indices in relation to blood pressure phenotypes and estimated pulse wave velocity (ePWV). We analyzed data from 29,294 adults in the National Health and Nutrition Examination Survey (NHANES) 1999–2018. Survey-weighted logistic and linear regression models examined associations of standardized TyG, TyG-body mass index (TyG-BMI), TyG-waist circumference (TyG-WC), and TyG-WWI with prehypertension, hypertension, abnormal blood pressure, systolic blood pressure (SBP), diastolic blood pressure (DBP), mean blood pressure (MBP), and ePWV. Restricted cubic splines assessed nonlinear associations, and unweighted receiver operating characteristic (ROC) curves with DeLong tests compared sample-based discrimination. In fully adjusted models, TyG-WC had the largest observed associations with categorical blood pressure phenotypes, including prehypertension (odds ratio [OR], 1.51), hypertension (OR, 1.83), and abnormal blood pressure (OR, 1.84) per standard deviation increase. TyG-WWI had the largest observed associations with SBP (beta, 3.12 mmHg) and ePWV (beta, 0.14 m/s); the ePWV result requires cautious interpretation because ePWV is derived from age and MBP. Restricted cubic splines showed nonlinear associations, with model-derived OR = 1.0 crossing points. For hypertension, TyG-WWI had the highest AUC (0.75), whereas TyG-WC had the highest AUCs for prehypertension (0.67) and abnormal blood pressure (0.74). The absolute AUC differences were small. Associations were more pronounced in younger participants and in those without diabetes. In this cross-sectional analysis, TyG-WC and TyG-WWI showed differential associations with blood pressure phenotypes and ePWV. TyG-WC had the largest observed associations with categorical blood pressure phenotypes, whereas TyG-WWI had a stronger association with SBP and the highest sample-based discrimination for hypertension. These indices may help characterize metabolic-anthropometric profiles associated with a higher prevalence of blood pressure abnormalities. Prospective studies with repeated blood pressure assessment, directly measured arterial stiffness, and external validation are needed before clinical risk-assessment use can be considered.