Association of the Triglyceride-Glucose Index and Its Central Obesity Derivatives with Peripheral Artery Disease in...

To investigate the associations of the triglyceride-glucose (TyG) index and TyG-related obesity indices with lower extremity peripheral artery disease (PAD) in type 2 diabetes mellitus (T2DM) patients, and to assess their discriminative ability.

This single-center, retrospective cross-sectional study included 2224 T2DM patients at a university-affiliated hospital between August 2020 and December 2024. PAD was defined as a resting ankle-brachial index (ABI) ≤ 0.90, or a post-exercise ABI decrease > 20%, or imaging-confirmed lower extremity arterial stenosis or thrombosis. Multivariable logistic regression, trend tests, subgroup analyses, and receiver operating characteristic (ROC) curves were applied.

Among the 2,224 patients (45.0% women, mean age 61.0 ± 11.3 years), the TyG index and its combinations with central obesity indicators (TyG-WHtR, TyG-WHR, TyG-WC) were positively associated with the presence of PAD (all P < 0.05), whereas the general obesity indicator (TyG-BMI) lost its significant association after multivariable adjustment, although it remained significant in specific subgroups, such as patients with well-controlled glycemia. After multivariable adjustment, patients in the highest quartile of these significant indices had 1.787-2.128 times higher odds of having PAD than those in the lowest quartile. Subgroup analyses identified significant interactions with sex and smoking status, showing stronger associations in women and non-smokers. Furthermore, ROC analyses indicated that TyG-WHtR had the highest discriminative ability (AUC = 0.622) among the indices, though its overall accuracy is modest.

In T2DM patients, the TyG index and its central obesity derivatives are independently associated with PAD, whereas TyG-BMI shows weaker relevance. TyG-WHtR demonstrates relatively better discriminative performance but limited accuracy, suggesting its role as a supplementary metabolic marker rather than a standalone screening tool.