Association of glycemic variability with short and long-term mortality among critically ill trauma patients: A retrospective study from the MIMIC-IV database

The relationship between glycemic variability (GV) and mortality in critically ill trauma patients is not well-established. We aimed to evaluate the association between GV, quantified by the coefficient of variation (CV), and both short- and long-term mortality in this population. A cohort of patients was established from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database. Cox proportional-hazards models, Kaplan-Meier analysis, restricted cubic splines (RCS), and subgroup analyses were used to investigate the association between GV and mortality. A mediation model was constructed to determine the mediating role of sepsis. This study included 4,009 critically ill trauma patients. Higher GV was independently associated with increased 30-day (Adjusted HR 1.49, P  < 0.001) and 1-year mortality (Adjusted HR 1.28, P  < 0.001). Nonlinear analyses revealed a J-shaped relationship, with mortality risk increasing sharply above a CV of 12.2%. The association was more pronounced in younger patients and those without diabetes. Mediation analysis revealed that sepsis significantly mediated this association, with proportions of 50.7% for 30-day and 70.5% for 1-year mortality. Higher glycemic variability is an independent predictor of both short- and long-term mortality in critically ill trauma patients. The risk appears to have a threshold effect, and sepsis is a major mediating pathway.