Glucose-to-albumin ratio and mortality risk in critically ill patients with ischemic stroke: A MIMIC-IV retrospective cohort study

The prognostic relevance of the admission glucose-to-albumin ratio (GAR) in critically ill patients with ischemic stroke (IS) has not been well established. We evaluated whether admission GAR is associated with subsequent all-cause mortality among intensive care unit patients with IS. Using the Medical Information Mart for Intensive Care IV database, we conducted a retrospective cohort analysis of adults admitted to the intensive care unit with a diagnosis of IS and available admission glucose and albumin. The primary endpoint was 90-day all-cause mortality; secondary endpoints included 30-day, 1-year, and in-hospital mortality. Patients were grouped by the median GAR. We used multivariable Cox regression, Kaplan–Meier curves, restricted cubic splines to evaluate dose–response patterns, and 1:1 propensity score matching to reduce baseline imbalance. Discrimination for mortality was assessed by receiver operating characteristic analysis. Among 2293 eligible patients, higher GAR was consistently associated with increased 90-day and 1-year mortality, and the association persisted after propensity score matching, with spline analyses suggesting an overall positive association between GAR and mortality risk. Using a clinically oriented cutoff, GAR ≥ 42.069 was associated with increased 90-day mortality (hazard ratio 1.27, P = .005). In critically ill IS patients, admission GAR independently predicts higher short- and long-term mortality and offers modest discrimination, suggesting potential utility for early risk stratification. External, prospective multicenter validation is needed.