Association between leuko-glycemic index and mortality in critically ill patients with non-traumatic subarachnoid hemorrhage: analysis of the MIMIC-IV database

Non-traumatic subarachnoid hemorrhage (NTSAH), primarily caused by intracranial aneurysm rupture, represents a significant global health challenge due to its high mortality and morbidity. The leuko-glycemic index (LGI), a biomarker reflecting oxidative stress and inflammation, has been associated with adverse cardiovascular outcomes. However, its prognostic value in critically ill NTSAH patients remains uncertain. Understanding the relationship between LGI and patient outcomes is essential to improve clinical management of NTSAH. We identified NTSAH patients from the Medical Information Mart for Intensive Care-IV (MIMIC-IV, version 2.2) database. Participants were divided into quartiles based on their LGI scores. Mortality was evaluated at multiple time points: ICU stay, in-hospital, and at 1-, 6-, and 12-month post-admission. The association between LGI and mortality was examined using multivariate Cox proportional hazards regression. Restricted cubic spline (RCS) analysis was employed to delineate the relationship between LGI scores and mortality risk and to identify the cutoff value. The robustness of these findings were confirmed through subgroup analyses, interaction tests, and likelihood ratio tests. A total of 750 patients were included, with 57% being female. Mortality rates were 17% in the ICU, 20% in-hospital, 21% at 1 month, 27% at 6 months, and 29% at 1 year. Multivariate Cox regression analysis revealed that higher LGI score were significantly associated with increased mortality at 1 month, 6 months, and 1 year. RCS analysis demonstrated a positive correlation between elevated LGI scores and mortality risk. LGI is significantly associated with mortality in critically ill NTSAH patients, suggesting its potential as a prognostic biomarker for risk stratification. Further validation through prospective cohort studies is necessary to confirm these findings.