Association between the triglycerides and glucose index and critically ill stroke in non-obese population: An observational study on eICU database.

Background: Both the triglyceride glucose (TyG) index and stroke are associated with insulin resistance (IR). Studies have shown that TyG is associated with stroke. However, the relationship between TyG index and stroke in non-obese population remains unclear. This study investigated the association between TyG and the prognosis of severe stroke in non-obese individuals.

Methods: This was a retrospective observational study. TyG was determined at baseline and stroke mortality was assessed at follow-up. All data were extracted from the eICU Collaborative Research Database. Hospital and intensive care unit (ICU) mortality rates were determined as endpoints. Smoothing curves were performed to estimate the relationship between TyG and stroke mortality. Multivariate logistic regression analysis and restricted cubic spline regression were used to evaluate the independent effect of TyG on stroke mortality.

Results: A total of 1946 eligible subjects were included. The hospital and ICU mortality rates were 10.6 % and 5.3 %, respectively. In univariate analysis, for per unit increase in TyG, the odds ratio (OR) for in-hospital mortality and ICU mortality of stroke were 1.85 (95 %CI 1.50-2.28) and 2.08 (95 %CI 1.58-2.75), respectively. After fully adjusting for confounding factors, the OR for in-hospital mortality and ICU mortality of stroke were 8.64 (95 %CI 4.25-17.55) and 9.30 (95 %CI 3.73-23.20), respectively. The restricted cubic spline regression model indicated a linear increase in hospital and ICU mortality with rising TyG index. Subgroup analysis showed consistent effect sizes and directions across different subgroups, demonstrating the stability of the findings.

Conclusion: In the non-obese population, an increase in TyG was related to higher stroke mortality. TyG may have clinical significance in identifying poor prognosis of stroke patients.