Subdural hematoma with mild symptoms progressed significantly less often when treated with middle meningeal artery (MMA) embolization compared with observation alone, new results from the second phase of the EMBOLISE trial showed.

In the prospective, multicenter randomized EMBOLISE trial, embolization of the MMA using the liquid embolic agent Onyx (Medtronic) reduced the composite primary endpoint of surgical drainage or clinical deterioration by 50% compared with conservative management in patients with mildly symptomatic subdural hematoma who were not candidates for surgery.

Onyx is a liquid embolic agent that is injected into an artery to occlude blood flow. In subdural hematoma, it is used to block the MMA, stopping the chronic oozing of blood into the subdural space.

The technique has demonstrated benefit in surgically treated subdural hematoma, and the current results show that it is also effective in nonsurgical patients.

“We are very excited to see a twofold reduction in the primary clinical endpoint in patients treated with embolization vs those managed conservatively,” said study investigator, Jared Knopman, MD, neurosurgeon and interventional neuroradiologist at Weill Cornell Medical College, New York City.

Knopman presented the latest results from the EMBOLISE trial on February 5 at International Stroke Conference (ISC) 2026.

Multicenter Trial

Conducted at 60 US centers, the trial evaluated the role of MMA embolization with Onyx in the management of chronic subdural hematomas.

Reported in 2024 the first part of trial showed a positive effect of Onyx in a surgical cohort of patients and led to FDA approval of Onyx for MMA embolization as an adjunct to surgical therapy.

The results from the second part of the trial, known as EMBOLISE-2 and presented at ISC 2026, focused on a more mildly symptomatic patient cohort who were not considered candidates for surgery.

“EMBOLISE-2 represents the first and only trial to date specifically powered to study this more mildly symptomatic patient population,” Knopman noted.

The trial included 200 patients with subdural hematoma with a thickness < 15 mm and < 5 mm of midline shift. The vast majority of patients had mild symptoms, predominantly headaches and gait instability.

Patients were randomly assigned to receive either observation or treatment with upfront MMA embolization with Onyx.

The two groups were well balanced in terms of baseline characteristics. The average subdural hematoma size was about 12 mm, with 2 mm of midline shift in each group.

There was a 27% loss to follow-up in the trial, which Knopman said was exacerbated by the COVID epidemic.

However, he noted that similar figures for loss to follow-up have been seen in other studies of nonsurgical patients with subdural hematoma.

“We are dealing with patients that are only mildly symptomatic and less likely to come back for follow-up,” he said.

He also pointed out that the primary endpoint in the trial was a composite and if any one of those endpoints was lacking, it would count as a loss to follow-up.

Number Needed to Treat of Nine

Primary results showed that the primary composite endpoint of either need for surgical drainage or poor clinical outcome or clinical deterioration at 90 days was halved in the group that underwent MMA embolization with Onyx (11.9%) vs 23% in the observation group (relative risk, 0.51; P = .04).

Poor clinical outcome was defined as a Markwalder score 2-4, motor strength score 0-3, or Glasgow Coma score 3-12; and clinical deterioration was defined as 1 or more point worsening in the Markwalder, motor strength, or Glasgow Coma score at 90 days.

The number needed to treat to prevent one primary endpoint adverse outcome was nine.

“This has vast implications for a disease state which affects an ever-growing number of our patient population and is growing in terms of deterioration of neurologic function,” Knopman said.

The number of patients experiencing deterioration of neurologic function did not meet noninferiority at 90 day (19.5% Onyx vs 9.7% control). However, noninferiority was met at 180 days (17% vs 15.2%).

Knopman said the finding was not surprising, noting that interventional studies often show higher early risk compared with conservative management, particularly when adverse events occur.

“What we did find surprising, though, is that the conservatively managed group did not have a benign course and that they paid the price down the line for consistently having that hematoma burden.”

“The takeaway here is that these patients are starting out with smaller subdural hematomas and may take a longer to manifest symptoms, but over time, they end up doing so.”

Hematoma volume was also reduced in the Onyx treated group at 50.3 ml vs 86.2 ml at 90 days and 43.8 ml vs 80.0 ml at 180 days.

“This again shows that those managed conservatively sustain a larger ongoing hematoma burden,” Knopman noted.

Adverse Outcomes Analyzed

Adverse events included a higher rate of stroke in the Onyx group (6%) compared with no strokes in the control group.

Knopman said that while the imbalance was “concerning,” further review showed that none of the strokes were directly attributable to the treatment. Three of the six strokes occurred in patients who had been taking antiplatelet or anticoagulation therapy that was withheld for the procedure.

Two others were related to catheterization of the internal carotid artery — one resulting in a dissection and the other in a shower of emboli. The remaining stroke was thought to be related to the patient’s underlying medical condition, with a very high Charlson Comorbidity Index (CCI) score indicating multiple comorbidities.

“The fact that five out of six of these strokes could be highly mitigated or prevented really taught us a couple of things about the procedure. One is that antithrombotic medication should not be stopped for secondary stroke prevention in this patient population, and the actual procedure itself should be kept simple, in the sense that you don’t want to perform potentially unnecessary catheterizations,” he said.

Neurologic death occurred in 4% of patients in the Onyx group compared with 1% in the control group, but Knopman said the deaths were not attributable to Onyx itself. Instead, they were strongly associated with CCI scores and the interruption of antithrombotic therapies.

Knopman described the medical management aspect of the study as particularly interesting, noting that there is no evidence to support stopping antiplatelet or anticoagulant therapy in nonsurgical patients. He said this was not surprising given that subdural hematoma progression is largely driven by inflammatory membrane leakage rather than acute re-bleeding.

“I’m probably preaching to the choir at a conference like ISC to stress the importance of secondary stroke prevention for our patients. But I think the neurosurgical community, which is a little bit more reflexive in stopping these agents, can learn something from what we’re finding with respect to this,” he said.

Knopman said that adhering to several key procedural principles — avoiding catheterization of the internal carotid artery in patients with difficult aortic arch anatomy, maintaining a lower threshold for aborting the procedure, and not discontinuing medications needed for secondary stroke prevention — could help reduce risk.

“We think that we can mitigate the risks and maximize the robust clinical effects that embolization has on primary endpoint reduction even in this mildly symptomatic patient population,” he said.

Need for Larger Studies

Commenting on the findings for Medscape Medical News, Bijoy Menon, MD, vice-chair of the ISC and a professor of neurology at the University of Calgary in Calgary, Alberta, Canada, said the study showed promise for MMA embolization with Onyx in patients with subdural hematoma who have milder symptoms and are not considered candidates for surgery.

“The results suggest that this technique could potentially help prevent surgery and neurological deterioration in the future.”

Menon said he chose the word “potentially” deliberately, noting that the study was relatively small and had some loss to follow-up. It is important, he said, to understand who these patients lost to follow-up were.

He said that larger studies with complete follow-up are needed to better inform treatment decisions in this patient population.

Medtronic sponsored the trial. Knopman reported being a consultant to Medtronic Neurovascular.