BACKGROUND: The triglyceride glucose-body mass index (TyG-BMI) has been established as a convenient and reliable marker for assessing insulin resistance (IR) and has been shown to be significantly correlated with stroke. However, only a few studies have been conducted in this field, with conflicting conclusions. METHODS: This study based on the eICU database, investigated the association between TyG-BMI and 28-day mortality in critically ill ischemic stroke (IS) patients. Multivariate Cox regression models were employed to analyze the impacts of the TyG-BMI on 28-day hospital and ICU mortality. Restricted cubic splines (RCS) were applied to explore the nonlinear relationship between the TyG-BMI and 28-day mortality. K‒M curves were utilized for outcome comparisons among different TyG-BMI groups. Additionally, interaction and subgroup analyses were performed to validate the robustness of the results. RESULTS: A total of 1,362 critically ill patients with IS were enrolled, with a mean age of 68.41 ± 14.16 years
47.50% were male. Multivariate Cox regression analysis revealed that, the high TyG-BMI group had significantly higher 28-day hospital mortality(HR = 1.734, P = 0.032) and ICU mortality (HR = 2.337, p = 0.048). RCS analysis showed a nonlinear positive correlation between the TyG-BMI and 28-day hospital mortality. Below the inflection point of the TyG-BMI = 380.37, each increase of 1 standard deviation (SD) (approximately 25.5 units) in the TyG-BMI was associated with a 37.3% increase in 28-day hospital mortality (HR = 1.373, P = 0.015), and above 380.376, each 1-SD increase in the TyG-BMI resulted in an 87.9% decrease in 28-day hospital mortality (HR = 0.121, P = 0.057). The log-likelihood ratio test P value = 0.004. For 28-day ICU mortality, the TyG-BMI exhibited a significant positive linear correlation in RCS. CONCLUSIONS: Elevated TyG-BMI is significantly associated with an increased risk of short-term all-cause mortality in patients with critically ill IS in the United States. This result provides compelling evidence to address the existing discrepancies in this research domain, indicating that the TyG-BMI could serve as a straightforward and efficient biomarker for identifying critically ill IS patients at high risk of mortality.