Abstract
Respiratory failure (RF) lead to high mortality rates and extended hospital stays in intensive care unit (ICU). The Triglyceride-Glucose (TyG) index, a reliable surrogate marker for insulin resistance (IR), predicted adverse outcomes in various diseases. Combining weight-related indices like body mass index (BMI) with TyG to form the TyG-BMI enhanced the assessment of IR and its impact on patient outcomes. However, the association between TyG-BMI and outcomes in patients with RF remained underexplored. This study retrospectively analyzed data from the MIMIC-IV database, focusing on critically ill patients with RF. From an initial cohort of 19,429 patients, 2177 met the inclusion criteria and were divided into quartiles based on TyG-BMI values. Key clinical information was collected within the first 24 h of ICU admission, including demographics, lab results, vital signs, and scoring systems such as SAPS II and SOFA. Primary outcome was 28-day, secondary outcomes were 180-day and 1-year mortality. Data were analyzed using multivariable Cox regression models, Kaplan-Meier survival curves, and restricted cubic splines to assess the nonlinear relationship between TyG-BMI and mortality. The study found significant differences in baseline characteristics across TyG-BMI quartiles. Kaplan-Meier survival curves indicated a higher survival probability for patients in the lowest TyG-BMI quartile (Q1) compared to higher quartiles (Q2-Q4). Adjusted hazard ratios demonstrated a nonlinear association between higher TyG-BMI values and increased mortality risk at all three time points. The RCS-derived cut-off value of 269 for TyG-BMI was identified as a significant threshold, with higher TyG-BMI values correlating with lower mortality risks. Subgroup analyses reinforced these findings across different patient demographics and clinical profiles. Higher TyG-BMI was associated to lower short-term and long-term mortality, suggesting a potential protective effect. These findings highlighted the importance of the TyG-BMI as a robust prognostic marker, providing valuable insights for improving treatment strategies for patients with RF.