Triglyceride-Glucose Index and Systemic Immune-Inflammation Index Predict All-Cause Mortality in Critically Ill Patients: A Retrospective Cohort Study From the MIMIC-IV3.1 Database

Background: Triglyceride-glucose (TyG) index is a biomarker of insulin resistance in metabolic diseases. Systemic immune-inflammation (SII) index quantifies inflammatory burden in conditions like sepsis and trauma. However, their individual and combined predictive value in intensive care unit (ICU) risk prediction remains underexplored. This study aimed to investigate the individual and combined associations of TyG and SII with the risk of all-cause mortality in ICU patients.

Methods: Data of this retrospective observational study were obtained from the Medical Information Mart for Intensive Care IV (MIMIC-IV-3.1) database. 2475 participants were stratified into tertiles based on the TyG index. The primary endpoint was 1-year all-cause mortality; the secondary endpoint was in-hospital mortality. Associations were assessed using multivariable-adjusted Cox proportional hazards and logistic regression models, restricted cubic splines (RCS) analysis, and receiver operating characteristic (ROC) curves to compare predictive performance against conventional severity scores (APS III, OASIS, CCI, SIRS).

Results: The 1-year and in-hospital all-cause mortality reached 30.1% and 16.9%, respectively. The cumulative incidence of 1-year all-cause death increased across higher tertiles of SII and TyG-SII within each TyG tertile group, while no significant differences were observed among the TyG tertiles themselves. RCS analysis revealed an approximately linear relationship between TyG index and in-hospital mortality. Both SII and TyG-SII were non-linearly related to 1-year and in-hospital all-cause mortality. After adjusting for confounding factors, patients in the highest TyG index group had a significantly higher risk of in-hospital mortality (OR 1.446 [95% CI 1.083-1.986]; P = 0.013) as well as 1-year all-cause mortality (HR 1.223 [95% CI 1.005-1.487]; P = 0.044) . Similar results were observed from SII and TyG-SII. The area under the ROC curve (AUC) for TyG or TyG-SII combined with conventional severity scores was higher than that for SII alone.

Conclusion: TyG, SII and TyG-SII are significantly associated with 1-year and in-hospital all-cause mortality. The combination of TyG or TyG-SII with traditional severity scores demonstrated superior predictive performance for clinical deaths of ICU patients compared to SII alone.

Keywords: Triglyceride-glucose index, Systemic immune-inflammation index, Intensive care unit, All-cause mortality, MIMIC-IV database

The Presenting Author is a PhD candidate specializing in the field of cardiovascular research. Engaged in studying the mechanisms of cardiovascular diseases, focusing on innovative therapeutic approaches and disease prevention. Involved in both experimental and clinical research to advance understanding of heart health and improve patient outcomes. Committed to contributing to the scientific community through publications and collaborative projects.