Abstract
OBJECTIVE: The triglyceride-glucose index (TyG) has been actively researched for predicting several diseases. However, high-quality evidence assessing its ability to predict stroke is lacking. We conducted a meta-analysis of high-quality studies examining if TyG can predict stroke in the general population. METHODS: Embase, PubMed, CENTRAL, Web of Science, and Scopus databases were searched until 13th January 2025. Cohort studies on the general population, excluding those with baseline stroke or cardiovascular disease, with a minimum follow-up of four years and reporting an adjusted association between TyG and stroke were included. TyG was assessed as both a categorical and continuous variable. RESULTS: A total of 13 studies with 12,898,434 individuals were eligible. The overall incidence of stroke was 0.89%. Meta-analysis indicated a statistically significant increased risk of stroke between higher vs lower values of TyG (risk ratio (RR): 1.27 95% confidence interval (CI) [1.19-1.35] I (2) = 66%). Per unit increase in TyG was also associated with a statistically significant increase in the risk of stroke (RR: 1.16 95% CI [1.07-1.27] I (2) = 89%). Most results remained unchanged on subgroup analysis based on location, excluded population, stroke diagnosis, TyG data, and follow-up. Meta-regression using moderators sample size, age, male gender, diabetes mellitus, hypertension, TyG cut-off, stroke incidence, and follow-up also failed to reveal significant results. CONCLUSION: High TyG is associated with increased risk of stroke in the general population.