Skip to content

Risk Score for Early Brain Injury Predicts 90-Day Outcome of Nontraumatic Subarachnoid Hemorrhage

Key findings

  • Treating early brain injury (EBI) after aneurysmal nontraumatic subarachnoid hemorrhage—the physiological imbalances that occur within 72 hours of symptom onset—may reduce neurological damage
  • In 190 patients, researchers at Massachusetts General Hospital derived an EBI risk score that considers age, sex, symptomatic hydrocephalus, Glasgow Coma Scale score 24 hours after stroke onset and the plasma level of soluble ST2
  • For predicting 90-day functional outcome (score on the modified Rankin scale), the new tool had an area under the receiving operating characteristic curve of 0.910, a sensitivity of 93%, specificity of 75%, and accuracy of 82%
  • The EBI risk score was validated in an independent cohort, highlighting its ability to perform across separate patient populations

After aneurysmal nontraumatic subarachnoid hemorrhage (SAH), brain injury from delayed cerebral ischemia and cerebral vasospasm negatively affects patient outcomes, but pharmacologic treatment of those consequences has been unsuccessful.

In exploring other ways to treat SAH, some scientists are focusing on early brain injury (EBI), defined as an injury that occurs within 72 hours, before the onset of vasospasm. Physiological derangements (e.g., rising intracranial pressure, decreased cerebral blood flow and global cerebral ischemia) are thought to lead to secondary injuries such as blood–brain barrier disruption, inflammation and oxidative cascades, and result in cell death.

Researchers at Massachusetts General Hospital have confirmed the importance of EBI to the long-term effects of SAH, and in the process, they have developed a risk score that can be used for patient selection or risk stratification in clinical trials of EBI therapies. Riana L. Schleicher, research coordinator, and W. Taylor Kimberly, MD, PhD, chief of the Division of Neurocritical Care in the Department of Neurology, and colleagues report the details in Stroke.

Developing a Risk Score

The team conducted a retrospective observational study of 190 Mass General patients with nontraumatic SAH (62% female; mean age 57). They collected data on clinical and radiographic markers of EBI:

  • Early loss of consciousness (at SAH onset)
  • Symptomatic hydrocephalus (a requirement for external ventricular drainage)
  • Glasgow Coma Scale (GCS) score closest to 24 hours from SAH onset
  • Early neuroworsening (decrease in GCS ≥2 points within the first 72 hours after SAH onset)
  • Global cerebral edema on radiography
  • Bicaudate index, a radiographic measure of hydrocephalus

In univariable analysis, all markers except cerebral edema predicted a poorer score on the modified Rankin Scale at 90 days.

At three time points (mean 3.5, 8, and 13 days), the researchers measured plasma levels of soluble ST2 (sST2), a marker of inflammation. sST2 levels were an independent predictor of 90-day outcome even after adjustment for 24-hour GCS and symptomatic hydrocephalus. Higher sST2 levels were also significantly associated with the early loss of consciousness and elevated bicaudate index.

The Final Model

The researchers conducted a multivariable analysis of all significant EBI markers plus sST2, adjusting for age, sex and severity of SAH on admission as rated on the Hunt–Hess Scale. Independent predictors of 90-day mRS score are listed below, along with the points they were assigned in the final model (total possible score=7):

  • Age—0 points for ≤60, 1 point for >60
  • Sex—0 points for female, 1 point for male
  • Symptomatic hydrocephalus—0 points for absence, 1 point for presence
  • 24-hour GCS score—0 points for >12, 2 points for ≤12
  • sST2 level—0 points for <66 ng/mL, 1 point for ≥66 ng/mL

The risk score had an area under the receiving operating characteristic curve of 0.910, a sensitivity of 93%, specificity of 75%, and accuracy of 82% for predicting 90-day functional outcome.

Validating the New Tool

The risk score was validated in an independent cohort of 50 patients with nontraumatic SAH from a hospital in Copenhagen (90% female; mean age 61). The area under the curve was 0.846, sensitivity was 86%, specificity was 83%, and accuracy was 84%.

sensitivity of a novel early brain injury score in predicting 90-day functional outcome after nontraumatic subarachnoid hemorrhage

accuracy of a novel early brain injury score in predicting 90-day functional outcome after nontraumatic subarachnoid hemorrhage

Learn about the Division of Neurocritical Care

Refer a patient to the Department of Neurosurgery

Related topics


Using two methods of machine learning, Matthew Koch, MD, Aman Patel, MD, W. Taylor Kimberly, MD, PhD. and colleagues have linked increased cerebrospinal fluid concentrations of three metabolites that alter the nitric oxide pathway to poor outcome after aneurysmal subarachnoid hemorrhage.


Clinicians in the Department of Neurology at Massachusetts General Hospital found that new or worsening epileptiform abnormalities on continuous EEG have a strong, lasting association with poor functional outcome after nontraumatic subarachnoid hemorrhage, suggesting they may guide therapy.