The impact of stress hyperglycemia ratio and glycemic variability on 1-year all-cause mortality in critically ill patients with intracerebral hemorrhage: a two-center retrospective study

Background Stress hyperglycemia ratio (SHR) and glycemic variability (GV) are dynamic markers of dysglycemia. This study evaluated their predictive value for mortality and their bidirectional mediation effect in critically ill patients with intracerebral hemorrhage (ICH). Methods This retrospective study included 1867 critically ill patients with ICH from the MIMIC-IV database and 326 from an external center (2018–2024). The primary outcome was 1-year mortality. Various statistical methods were employed, including multivariate Cox proportional hazards regression, Kaplan-Meier analysis, and restricted cubic splines (RCS). Results Ultimately, 1867 critically ill patients with ICH were extracted from the MIMIC-IV database. The 1-year and in-hospital mortality rates were 21.75% and 14.78%, respectively. RCS analysis revealed a nonlinear association between SHR and 1-year all-cause mortality, and a linear association for GV. Among non-diabetic patients, those with high SHR and high GV levels had the highest risk of both in-hospital and 1-year mortality. However, among diabetic patients, those with high SHR and low GV levels had the highest risk of both in-hospital and 1-year mortality. In addition, receiver operating characteristic (ROC) curve analysis showed that the combination of SHR and GV outperformed either measure alone in predicting mortality. Finally, all results were further confirmed in an external cohort. Conclusions Our study demonstrates that SHR and GV are significantly associated with 1-year mortality in critically ill patients with ICH. The combined assessment of SHR and GV is crucial for identifying at-risk individuals, optimizing treatment strategies, and improving patient outcomes.