Enhanced prediction of in-hospital mortality in intracerebral hemorrhage: impact of serial neurological and radiological reassessment with the ICH Score at 6 hours postadmission

OBJECTIVE The authors of this study aimed to identify predictors of in-hospital mortality in patients with primary supratentorial intracerebral hemorrhage (ICH) at emergency department admission and 6 hours thereafter. Additionally, they evaluated the predictive accuracy of a modified ICH (mICH) Score incorporating midline shift (MLS), compared to that of the original ICH Score. METHODS This retrospective analysis included adult patients with primary supratentorial ICH who had been admitted to a Comprehensive Stroke Center between July 2017 and December 2023. Data extracted from the electronic medical records included demographics, clinical history, blood pressure, ICH characteristics on CT scans (i.e., location, hematoma volume, intraventricular hemorrhage, MLS), Glasgow Coma Scale (GCS) score, ICH Score, laboratory tests (i.e., white blood cell WBC count and hemoglobin, hematocrit, platelet, and glucose levels), antithrombotic use, neurological interventions, and discharge status. The primary outcome was in-hospital mortality. The mICH Score was calculated by substituting ICH volume in the original risk stratification scale with MLS (≥ 5 mm = 1 point). Statistical analyses included descriptive statistics, chi-square test, t-test, logistic regression, and receiver operating characteristic curve analysis. RESULTS The in-hospital mortality rate among 518 patients with primary supratentorial ICH was 23%. Compared with survivors, deceased patients were older, had lower BMIs, more frequently presented with loss of consciousness, and had lower GCS scores and higher ICH Scores at admission and 6 hours thereafter. Independent predictors of death included older age, lower BMI, cortical ICH location, hematoma volume ≥ 30 cm 3 , intraventricular hemorrhage, MLS ≥ 5 mm, lower GCS score, higher ICH Score, elevated systolic blood pressure, higher WBC count and glucose level, and lower hemoglobin and hematocrit levels. On admission, the ICH Score (area under the curve AUC 0.890) and GCS score (AUC 0.879) showed a strong predictive performance for mortality, which improved at 6 hours after admission (AUC 0.914 for both). The mICH Score (AUC 0.897) demonstrated predictive accuracy comparable to that of the ICH Score. Twenty-one percent of the patients experienced ICH Score progression at 6 hours, which was associated with a 2.4-fold increase in mortality risk. CONCLUSIONS Findings in this study confirm established predictors of mortality in supratentorial ICH and highlight the prognostic value of neurological assessment 6 hours after admission. The mICH Score offers a practical and similarly accurate alternative to the original ICH Score for predicting in-hospital mortality. These findings underscore the importance of early and serial assessments to guide risk stratification in patients with ICH.