Maximum stress hyperglycemia ratio predicts mortality in critical hemorrhagic stroke: a retrospective cohort study from MIMIC-IV

Background : The stress hyperglycemia ratio (SHR), which quantifies acute glucose fluctuations relative to chronic glucose control, has been understudied in hemorrhagic stroke (HS). This study aimed to evaluate the prognostic utility of maximum SHR in predicting mortality in critically ill HS patients. Methods : We extracted 1,677 HS critically ill patients from the MIMIC-IV database and analysed them retrospectively. Logistic regression was used to assess the relationship between maximum SHR and mortality, and Kaplan-Meier and restricted cubic spline curves were plotted. Result : Among 1,677 included patients (median age 70.5 years, 53.8% male), logistic regression analyses showed that maximum SHR was associated with in- hospital mortality (per SD increase: OR 1.56, 95% CI 1.33-1.83), ICU mortality (per SD increase: OR 1.64, 95% CI 1.39-1.95), and 1-year mortality (per SD increase: OR 1.27, 95% CI 1.11-1.45). The correlation was approximately linear. Sensitivity tests demonstrated the consistency of the outcomes. Furthermore, integrating the maximum SHR into standard models enhanced their ability to discriminate and reclassify. Conclusion : A significant link was observed between the maximum SHR and the mortality of HS patients in the ICU. Incorporating the maximum SHR into current models enhances their forecasting accuracy, proving clinically important in pinpointing patients at high risk.