Clinical and neuroimaging predictors of early hospital mortality in patients with hemorrhagic stroke

Abstract

Objective: To date the complex interplay of predictors of stroke-associated mortality have been stated. Despite the abundancy of literature data on stroke-associated mortality, little is known about the predictors of early post-stroke mortality in developing countries. In this study, we aimed to analyze predictors of in-hospital stroke-associated mortality in Kyrgyz patients. Methods: Research design – retrospective, cross-sectional. Retrospective findings from medical records of 64 patients with hemorrhagic stroke (HS) were used for study. Collected data included clinical, laboratory, neuroimaging parameters, as well as autopsy protocols, which could be associated with a stroke-associated mortality in patients hospitalized in the tertiary care center with distinct stroke department. All patients were divided into 2 groups: deceased and survived ones. Initially comparison tests were applied for groups. Then multiple regression analysis was conducted to determine predictors of stroke associated mortality. Results: The study included patients with the following clinical subtypes: 12 patients were with subarachnoid hemorrhage (SAH), 42 patients with intracerebral hemorrhage (ICH) and 10 patients with combined (SAH+ICH) HS. Our study showed that the severe disorder of consciousness (coma) (66%-5.7%, p<0.001), impaired functioning of vital organs (hypoxemia 62%-11.4% - p<0.001; tachycardia – 51%-14.3% - p<0.006), the need for ventilator support and catheterization of the bladder (100% - 0% - p<0.001), neuroimaging picture of threatening cerebral edema with signs of its insertion (50%-20% - p<0.006) and expansion of hematoma into the ventricular system (60%-20% - p<0.006), hyperglycemia (65%-16.7% - p<0.001) and hyperthermia (55%-2.9% - p<0.001), as well as impaired kidney function (65%-16.7% - p<0.001) were significantly more common in mortality groups than in survived patients. Multiple regression analysis demonstrated that only systolic blood pressure level was associated with better survival (OR – 0.04, 95%CI 0.93-0.99, p=0.01). Conclusions: Thus, as our research shows, that severe unconsciousness, the expansion of hematoma into the ventricular system, progressive cerebral edema with dislocation of medial structures and impairment of the brain, multi-organ insufficiency, tachypnoe/bradypnoe, tachycardia/bradycardia, diabetes mellitus and increased level of cholesterol were more common in HS patients with in-hospital mortality as compared to survivors. The only significant predictor of in-hospital stroke associated mortality among clinical and laboratory variables was systolic blood pressure at onset, low oxygen saturation had borderline significance.

Keywords

• hemorrhagic stroke
• retrospective study
• predictors
• fatal outcome
• hospital stage
• risk stratification