Nancy A. Melville

October 18, 2016

BALTIMORE — Key clinical and radiographic characteristics from computed tomography (CT) assessed within 24 hours of ischemic stroke factor importantly in a clinical risk prediction model for patients at risk for potentially lethal malignant edema (PLME), new research suggests.

Malignant edema, which occurs in up to 10% of all strokes and in up to 50% of large hemispheric infarctions, is a serious complication, resulting in mortality in as many as 50% to 80% of cases.

While decompressive hemicraniectomy can substantially reduce mortality, early recognition of the need for surgery is critical for the intervention to have benefit.

"Our challenge is that clinical and even radiographic variables at baseline may not be sufficient to identify these patients," first author Charlene J. Ong, MD, MPHS, a neurocritical care fellow at Massachusetts General and Brigham and Women's Hospitals, Boston, said in presenting the research here at the American Neurological Association (ANA) 2016 Annual Meeting.

"Many of us have seen patients with large strokes who present early in their course, but it may not yet be clear whether they will develop malignant edema," she said.

Dr Charlene J. Ong

For this analysis, Dr Ong and colleagues identified 222 ischemic stroke patients presenting at Barnes-Jewish Hospital at Washington University School of Medicine, St. Louis, Missouri, between 2006 and 2015 with diagnoses of cerebral edema.

All patients had acute anterior circulation stroke with a National Institutes of Health Stroke Scale (NIHSS) score of 8 or higher and a baseline brain CT within 24 hours of stroke onset.

Among the patients, 73 developed PLME, including 50 who died of cerebral edema and 23 who had a hemicraniectomy and survived. Patients had a median age of 68, 60% were female, and the median NIHSS score was 18.

After multivariate adjustment, factors showing the strongest independent association with PLME included glucose (odds ratio [OR], 1.005; P = .016); previous stroke (which was protective; OR, 0.495; P = .01); no acute intervention, including tissue plasminogen activator or thrombectomy (OR, 1.940; P = .048); midline shift (OR, 1.29; P < .005); and baseline cistern effacement (OR, 5.273; P = .059).

Using data from the model, the authors developed an integer-based risk prediction scale, assigning value to each risk factor, with a score greater than 7 having a positive predictive value of 90%.

Strengths of the research include the study's size and clinical applicability, Dr Ong told Medscape Medical News. "This is the largest cohort in the literature that we are aware of, including 222 patients with cerebral edema," she said.

"Additionally, it's clinically oriented, with an applicable score for use for triage, as well as to help with surgical management and to facilitate communication."

While various other risk prediction scales exist, the new model provides a useful tool in settings where MRI is not readily available, Dr Ong noted.

"Of the few risk scores in the literature, some use MRI imaging data, but they're not necessarily generalizable to many centers around the US who may not have access to that imaging, which can be very labor and cost intensive."

"Others apply to patients arriving within 6 hours of onset, but that actually doesn't apply to up to 50% of patients who arrive with stroke," she noted.

"We hope to be able to help physicians and other ancillary staff in settings where they may not be used to seeing this population quite as much. The risk prediction scale can help in the management and decision making, as well as communication facilitation between healthcare professionals and family members when patients come in with this dire complication and need to figure out what to do."

The authors say the next step in the research is to externally validate the model with another cohort.

The study offers valuable information in the critical realm of cerebral edema after stroke, Kevin N. Sheth, MD, chief of the Division of Neurocritical Care and Emergency Neurology at Yale School of Medicine and Yale New Haven Hospital, Connecticut, told Medscape Medical News.

"The area is very important," he said. "We urgently need improved risk prediction models to facilitate the translation of novel therapies and for triage decisions."

Among caveats in the findings is the role of midline shift data, he noted.

"One of the challenges is the role of the midline shift data as far out as 24 hours," Dr Sheth said. "If the midline shift is already great, you probably don't need a risk prediction score. Things are already bad."

The authors have disclosed no relevant financial relationships. Dr Sheth is the principal investigator for a translational trial testing new molecules to prevent brain edema but had no disclosures directly relating to the study.

American Neurological Association (ANA) 2016 Annual Meeting. Abstract S219. Presented October 17, 2016.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: