Related Experiment Video
Updated: Dec 2, 2025
Minimally Invasive Endoscopic Intracerebral Hemorrhage Evacuation
Published on: October 15, 2021
Revised intracerebral hemorrhage expansion definitions: Relationship with care limitations
Ronda Lun1, Vignan Yogendrakumar1, Greg Walker1,2
1Ottawa Stroke Program, Department of Medicine (Neurology), University of Ottawa and Ottawa Hospital Research Institute, Ottawa, ON, Canada.
Background:
Hematoma expansion is an important therapeutic target in intracerebral hemorrhage. Recently proposed hematoma expansion definitions have not been validated, and no previous definition has accounted for withdrawal of care.
Aims:
To externally validate revised definitions of hematoma expansion that incorporate intraventricular hemorrhage, and to test their validity in the context of withdrawal of care.
Methods:
We analyzed data from the Antihypertensive Treatment of Acute Cerebral Hemorrhage II trial, comparing revised definitions of hematoma expansion incorporating intraventricular hemorrhage expansion to the conventional definition of "≥6 mL or ≥33%." Primary outcome was modified Rankin Scale of 4-6 at 90 days. We calculated the incidence, sensitivity, specificity, positive and negative predictive values, and c-statistic for all definitions of hematoma expansion. Definitions were compared using nonparametric methods. Secondary analyses were performed after removing patients with withdrawal of care.
Results:
Primary analysis included 948 patients. Using the conventional definition, the sensitivity was 37.1% and specificity was 83.2% for the primary outcome. Sensitivity improved with all three revised definitions (53.3%, 48.7%, and 45.3%, respectively), with minimal change to specificity (78.4%, 80.5%, and 81.0%, respectively). The greatest improvement was seen with the definition "≥6 mL or ≥33% or any intraventricular hemorrhage," with increased c-statistic from 60.2% to 65.9% (p < 0.001). Secondary analysis excluded 46 participants who experienced withdrawal of care. The revised definitions similarly outperformed the conventional definition in this population, with the greatest improvement in c-statistic using "≥6 mL or ≥33% or any intraventricular hemorrhage" (58.1% vs. 64.1%, p < 0.001).
Conclusions:
Revised hematoma expansion definitions incorporating intraventricular hemorrhage expansion outperformed conventional definitions for predicting poor outcome, even after accounting for care limitations.
