Association of various insulin resistance surrogate markers with mortality risk in critically ill patients with ischemic stroke: a retrospective cohort study

BACKGROUND: Ischemic stroke poses a substantial global health burden and represents one of the leading causes of death and disability. Insulin resistance (IR), a core feature of metabolic dysfunction, is recognized to play a critical role in the pathogenesis and progression of ischemic stroke. However, the predictive value of different surrogate markers of IR for mortality in patients with ischemic stroke remains unclear. METHODS: Using the public eICU-CRD database, we screened 1709 critically ill patients with ischemic stroke admitted to the intensive care unit (ICU). Patients were stratified by quartiles according to 8 distinct IR surrogate markers. The primary endpoint was in-hospital mortality. Statistical analyses included Cox regression, Kaplan-Meier curves, and restricted cubic spline (RCS) models. RESULTS: Among the 1709 ICU-admitted patients, 437 deaths occurred, with an in-hospital mortality rate of 25.57%. Multivariable Cox regression analyses demonstrated that all IR surrogate markers were significantly associated with mortality risk in patients with ischemic stroke. TyG-derived indices, TG-HDL, METS-IR, AIP, and CHG were positively correlated with mortality, whereas SPISE showed an inverse association (all P for trend < 0.05). Kaplan-Meier curves confirmed the associations between the 8 IR surrogate markers and in-hospital mortality. RCS analyses revealed nonlinear relationships between SPISE, TG-HDL, METS-IR, TyG-BMI, TyG-RC, CHG and all-cause mortality. CONCLUSION: Multiple IR surrogate markers independently predict in-hospital mortality in critically ill patients with ischemic stroke. Incorporation of these indices into risk stratification may facilitate early identification and targeted intervention for high-risk individuals.