Association of variability of blood pressure and glycemic with the risk of mortality and severe consciousness disturbance among critically ill patients with ischemic stroke, subarachnoid hemorrhage, and intracerebral hemorrhage: a retrospective cohort study from MIMIC-IV database

Background: Critically ill patients with stroke remains a major public health issue worldwide. Glycemic variability (GV) and systolic blood pressure variability (SBPV) are recognized as independent predictors of cardiovascular and cerebrovascular disease outcomes. This study aimed to explore the associations between GV and SBPV levels on the risk of in-hospital mortality, 1-year mortality and severe consciousness disturbance in different types of stroke patients including ischemic stroke (IS), subarachnoid hemorrhage (SAH), and intracerebral hemorrhage (ICH), and to further examine whether GV and SBPV exhibit interactive effects on these clinical outcomes. Methods: Data of this retrospective cohort study were extracted from the Medical Information Market for Intensive Care (MIMIC-IV) database. Severe consciousness disturbance was defined as Glasgow Coma Scale (GCS) points <8. Univariate and multivariate Cox proportional hazard models, logistics regression models, Kaplan-Meier (KM) analysis and restricted cubic splines (RCS) analysis were utilized to explore the associations of GV, SBPV, and their interactive effects on the risk of in-hospital mortality, 1-year mortality and severe consciousness disturbance, with hazard ratio (HR), odd ratio (OR) and 95% confidence interval (CI). Subgroup analyses were conducted based on different types of strokes. Results: Totally, 1,540 eligible patients were included, among them, 323 occurred in-hospital death, 523 occurred 1-year mortality and 463 occurred severe consciousness disturbance within 30-day after admission. High GV and high SBPV levels were associated with high in-hospital mortality risk (SBPV, HR=1.92, 95%CI: 1.41-2.59; GV, HR=1.84, 95%CI: 1.34-2.54) and high 1-year mortality risk (SBPV, HR=1.61, 95%CI: 1.24-2.08; GV, HR=1.57, 95%CI: 1.20-1.01). No significant associations were found between GV and SBPV with the risk of severe consciousness disturbance. High GV and high SBPV were exhibited a potential interactive effect on the risk of these outcomes we interested. We further observed that GV and SBPV demonstrated a significant interaction effect on the risk of in-hospital mortality (HR=and 6.80, 95%CI: 2.88-16.05, P <0.001) and 1-year mortality (HR=5.10, 95%CI: 2.54-10.27, P <0.001) among critically ill patients with ICH. The results of Kaplan-Meier analysis and RCS analysis demonstrated trends consistent with those observed in the multivariable Cox proportional hazard models. Conclusion: Our study suggested that in critically ill patients with IS, SAH and ICH, higher GV and SBPV levels were associated with higher risk of mortality. Furthermore, GV and SBPV exhibit interactive effects on both clinical prognosis and the development of severe consciousness disturbance in our study population. GV and SBPV can support doctors in identifying patients with high risk of mortality and making timely clinical decisions.