Triglyceride–Glucose Index Modifies Mortality Risk Across Body Mass Index Strata in Critically Ill Patients: A Retrospective Cohort Analysis of the MIMIC-IV Database

Background: In critically ill patients, the non-linear relationship between body mass index (BMI) and survival has persisted across studies, but whether it is affected by the triglyceride–glucose (TyG) index is unclear. Methods: We extracted critically ill patients without diabetes mellitus from the Medical Information Mart for Intensive Care (MIMIC)-IV database and categorized them into T1, T2, and T3 groups according to the TyG index tertiles, with the primary outcome of 365-day all-cause mortality. We used Kaplan–Meier (KM) survival curves, COX regression analyses, and restricted cubic spline curves (RCS) to assess the effect of TyG on the risk of BMI-related mortality. Estimation of optimal change points using segmented linear models. Results: The final analytic cohort comprised 6933 critically ill patients with a median age of 66 (54, 76) years, showing male predominance (60.9%). KM curves showed a lower protective effect of obesity in the high-TyG group. COX regression analyses showed that overweight and obesity were associated with lower 365-day all-cause mortality in the T1 and T2 groups and not in the T3 group. RCS analyses showed a U-shaped association between BMI and 365-day all-cause mortality in the T1 and T2 groups and a J-shaped association in the T3 group. The optimal BMI change points were 33 kg/m2 (T1), 32.5 kg/m2 (T2) and 29.7 kg/m2 (T3). Conclusions: Our findings demonstrate that the protective effect of obesity was significantly reduced in patients with high TyG levels. Consequently, it is necessary to consider the TyG index in the weight management of critically ill patients.