Visit-to-visit variability and longitudinal change of the triglyceride-glucose index are associated with early carotid plaque progression

The triglyceride-glucose (TyG) index is a surrogate marker of insulin resistance, but the clinical relevance of its longitudinal dynamics remains unclear. This study aimed to evaluate whether visit-to-visit variability in TyG (TyG-CV) and longitudinal change in TyG (TyG-Slope) were associated with early carotid plaque progression and improved risk prediction beyond conventional factors and mean TyG. In this single-centre retrospective longitudinal cohort, 1,519 adults aged 18–65 years who completed three annual examinations from 2023 to 2025 were included. Fasting triglycerides and fasting plasma glucose were measured from overnight fasting venous blood samples using standardized laboratory assays, and the TyG index was calculated at each visit. Carotid plaque and plaque area were assessed by standardized carotid ultrasonography. TyG-CV was calculated as SD/mean × 100%, and TyG-Slope was estimated as the annualized linear change in TyG across the three examinations. Plaque progression was defined as incident plaque in 2025 or an increase in total plaque area of ≥ 0.05 cm² from 2023 to 2025. Plaque progression occurred in 384 participants (25.3%) and was more frequent in higher TyG-CV quartiles. In the fully adjusted model, both TyG-CV and TyG-Slope were independently associated with plaque progression (per 1-SD increase: OR 1.28, 95% CI 1.13–1.45; and OR 1.26, 95% CI 1.12–1.42, respectively), whereas mean TyG was not. Restricted cubic spline analyses showed significant overall associations for both dynamic metrics, without statistically significant evidence of non-linearity. Adding spline-modelled TyG-CV or TyG-Slope was associated with modest improvements in discrimination and reclassification in internal model evaluation, while the model including both dynamic metrics achieved the highest predictive performance among the evaluated models. TyG-CV and TyG-Slope were independently associated with early carotid plaque progression and provided incremental predictive information beyond conventional factors and mean TyG in this single-centre health-screening cohort. External validation is needed before clinical application.