Subscribe to eTOC


SAN DIEGO—Intracranial hemorrhages are less severe in patients with untreated brain arteriovenous malformations (AVMs) than in patients with hemorrhages due to other causes, according to investigators who presented their findings here at the American Neurological Association Annual Meeting.

“The common perception is that these hemorrhages are as damaging as those due to hypertension or aneurysm,” said Jae H. Choi, MD, in an interview following his presentation. “We found that hemorrhages due to brain AVMs were mild in comparison, with a mean Rankin scale of zero to 1 [indicating no symptoms or significant disability] in 50 percent of cases.” Dr. Choi is a stroke fellow at the Neurological Institute of Columbia University Medical Center in New York City.


Typical site of arteriovenous malformation.


He and his co-investigators obtained data on 337 patients with cerebral AVMs from the Prospective Columbia AVM Databank. Among these, 152 (45 percent) had intracranial hemorrhage as the first manifestation. Other symptoms typically include headache, seizure, and stroke-like episodes, he added.

The investigators reviewed results from 132 who were evaluated within 30 days after the hemorrhage, comparing their records with 84 survivors of primary cerebral hemorrhage due to other causes. Those patients' records were drawn from the Northern Manhattan Stroke Study. The investigators also compared the effect of the hemorrhage's anatomic location: parenchymatous, subarachnoid, or ventricular. They assessed the severity of outcome with the Rankin scale and the National Institute of Health Stroke Scale (NIH-SS) score, an 11-item scale that assesses neurologic outcome and severity of disability.

The Rankin scale uses scores of 0 (for no symptoms) to 6 (for death). On the NIH-SS, a score of 15 or more indicates severe impairment, while 0 indicates no neurologic deficits post-stroke.

In this study, the investigators defined neurologic deficits by overall NIH-SS score as mild (with a score of 0 to 5); moderate (a score of 6 to 13), and severe (a score greater than 13).

Patients with AVMs were an average of 38 years old and equally divided between men and women; among 131 survivors – one patient with a primary hemorrhage died – the median Rankin scale score was 2, indicating slight disability. Twenty-four patients (18 percent) had a recurrent hemorrhage and one of them died. The average NIH-SS score of 5.9 (in the mild range) was significantly lower in the AVM patients, than the 13.6 (for more severe deficits) reported in the Northern Manhattan Stroke Study with p equal to or less than 0.0001.

In the AVM patients, 73 percent had an NIH-SS score of zero to 5; 17 percent had a score of 6 to 13, and 5 percent had a score of greater than 13. In the Northern Manhattan Stroke Study group, 24 percent had an NIH-SS score of zero to 5, 33 percent had a score of 6 to 13, and 43 percent had a score of greater than 13.

In both groups, the anatomic location affected the outcomes. For example, parenchymal hemorrhages were associated with higher NIH-SS scores than other bleeding sites, with a probability of 0.036.


Dr. Choi noted that cerebral AVMs are rare. They occur at a prevalence of three per 100,000, and an incidence of three per 100,000 per year, he said. “The lesions are usually first manifest by hemorrhage.” Historically, 50 percent of these AVMs have first symptoms with hemorrhage, he said, according to a review he and Jay P. Mohr, MD, published in The Lancet Neurology. Dr. Mohr is the Director of the Tananbaum Stroke Center at Columbia.

Dr. Choi stressed that the important finding of the study was that AVMs are not as devastating as hemorrhages due to hypertension or aneurysmal rupture. “We don't know about the comparative mortality of hemorrhages from brain AVMs and other causes,” he said, but morbidity in this study was mild.

He added that some of the patients studied had higher Rankin scores – about 14 to 20 percent – depending on the location of the lesion; parenchymal bleeds were associated with higher scores.

Dr. Choi also pointed out that the study had the typical limitations of a one-center study. To put things in perspective, he said, “This is one center reporting and comparing the results. Other centers may have patients whose course may differ. But it's a huge population and baseline characteristics were widely spread, so there were not only small AVMs or only young patients.”


“The findings are provocative, but I'm surprised by the results,” said Eelco FM Wijdicks, MD, who commented on the study in a phone interview. “The NIH-SS scores were below 5 in 73 percent of patients. I'm not sure that this accurately represents this group of patients.” Dr. Wijdicks, who was not involved in the study, is Professor of Neurology at Mayo Clinic in Rochester, MN, and Chair of the Division of Critical Care Neurology. He noted, though, that his perspective is informed by his experience with patients in the intensive care unit.

“Brain AVMs can be severe, as can bleeds in patients whose AVMs appear largely obliterated after treatment,” he said, referring to one method of treating the lesions. “I'm truly struck by how devastating first hemorrhages due to AVMs can be, but I am well aware I see only part of the spectrum.” Treatment options consist of embolization, radiation, and surgery, depending on the clinical features of the lesion.

“We can do so much in the treatment of AVMs, but we know so little about the natural history of brain AVMs,” Dr. Wijdicks said. “There may be too many variables to make conclusions about untreated AVMs that apply uniformly to the group.”


In upcoming research, the investigators hope to reduce some of that mystery by comparing the outcomes of untreated and treated unruptured AVMs, said Dr. Choi. In that study, A Randomized multicenter clinical trial of Unruptured Brain AVMs (ARUBA), the investigators will assess the outcome of patients with unruptured brain AVMs in those who undergo invasive treatment and those being clinically observed. The NIH-funded study, led by Dr. Mohr, will involve more than 80 centers and 800 patients, Dr. Choi said. The interventions under study will include endovascular embolization with glue, sterotactic radiotherapy, or surgical excision.


  • ✓ A new study reports that cerebral arteriovenous are not as devastating as hemorrhages due to hypertension or aneurysmal rupture.


• Choi JH, Mohr JP. Brain arteriovenous malformations in adults. Lancet Neurol 2005;4(5):299–308.