Nonlinear Association Between Triglyceride-Glucose Index and 28day Mortality in Intensive Care Units：A Multi-Center Retrospective Cohort Study

Bo Li ¹ Xiaoan Yang ² jiang hua Wang ¹ Wei dong Chen ¹ Qi Wang ^1* Lintao Zhong ^1*

• ¹ South China University of Technology, Guangzhou, Guangdong Province, China
• ² Third Affiliated Hospital of Sun Yat-sen University, Guangzhou, China

The triglyceride-glucose (TyG) index, derived from the calculation of two biomarkers, fasting plasma glucose and triglyceride levels, is a reliable indicator of insulin resistance and has been demonstrated to be associated with the adverse clinical outcomes of patients in the intensive care unit (ICU). This study aims to investigate the relationship between the TyG index and the 28-day all-cause mortality of these patients during their ICU stay.Methods: This study employed a multicenter retrospective cohort design, analyzing data from 18,883 ICU patients in the eICU database. We calculated the TyG index for each patient and assessed its association with 28-day all-cause mortality. The Cox proportional hazards model was utilized for analysis, adjusting for various clinical and laboratory variables to control for confounding factors. We performed sensitivity analyses, subgroup analyses, and interaction analyses to evaluate the robustness of the results.The study identified a significant positive correlation between the TyG index and 28-day all-cause mortality. Specifically, each one-unit increase in the TyG index corresponded to a 58% increase in mortality risk (HR=1.58, 95% CI: 1.25-2.00, P=0.0001). Additionally, the analysis revealed a non-linear threshold effect of the TyG index on mortality, with a cutoff point at 8.82; mortality was lower below this value and significantly increased above it. Sensitivity and subgroup analyses indicated robust findings, while E-value analysis suggested resilience against unmeasured confounding.This study establishes the TyG index as an independent predictor of 28-day all-cause mortality in critically ill patients, highlighting its potential value in clinical management and risk assessment. By recognizing the non-linear effect of the TyG index, clinicians can more effectively adjust treatment strategies to reduce mortality among high-risk patients.