Association of glycemic variability with short- and long-term mortality in critically ill patients with trauma.

The relationship between glycemic variability (GV) and mortality in critically ill trauma patients remains unclear. We evaluated whether GV, quantified by the coefficient of variation (CV) of glucose during the ICU stay, is associated with short- and long-term mortality in this population. Using the Medical Information Mart for Intensive Care IV (MIMIC-IV) database, we identified 4,009 adult trauma patients admitted to the intensive care unit and assessed the association between CV and 30-day and 1-year mortality with Cox proportional hazards models, Kaplan-Meier analysis, restricted cubic splines, subgroup analyses, and mediation models to explore the potential mediating role of sepsis. Higher GV was independently associated with increased 30-day (adjusted HR 1.50, P < 0.001) and 1-year mortality (adjusted HR 1.29, P < 0.001). Nonlinear analyses showed a J-shaped relationship between CV and mortality, with risk rising steeply above a CV of approximately 12.2%, particularly in younger patients and those without diabetes. Mediation analyses suggested that sepsis accounted for about 50.7% of the association with 30-day mortality and 70.5% with 1-year mortality. In critically ill trauma patients, higher GV is independently associated with both short- and long-term mortality, with an apparent threshold effect at mid-teen CV values, and sepsis appears to be a key pathway linking GV to mortality, although these mediation findings should be interpreted as exploratory and hypothesis-generating.