Diogo C. Haussen, MD


Bruce E. Pollock et al. The Risk of Stroke or Clinical Impairment After Stereotactic Radiosurgery for ARUBA-Eligible Patients. Stroke. 2013; 44: 437-441.

The initial experiences of surgical treatment for brain arteriovenous malformations (BAVM) in the early 1900s by Harvey Cushing and contemporaries were extremely challenging. Despite the remarkable advances in the understanding of this disease over the last decades, the most appropriate approach to unruptured BAVM remains unknown.

The ongoing ARUBA trial is a RCT of medical therapy vs. intervention (radiotherapy, neurosurgery and/or endovascular embolization). When completed, it will help us decide the best treatment for our patients with BAVM. Until the results of this trial are published, however, we must rely on other studies to guide our treatments. 

Pollock and colleagues report such a study in Stroke. In their interesting and timely study they analyzed the risks to patients undergoing stereotactic radiosurgery (SRS) for unruptured BAVM. They used the same eligibility criteria as the ARUBA trial to select retrospectively from their database 165 patients with 2 or more years of follow-up, They looked at the same outcomes as  ARUBA: the composite of ‘symptomatic stroke or death’, as well as ‘death or clinical impairment (modified Rankin Scale ≥2)’. 

The BAVM obliteration rate post initial SRS was ~70% and the time to obliteration was ~3.3 years. Fifteen patients (8.6%) had a hemorrhagic stroke during follow up and 4% developed permanent radiation-related focal neurological deficit after SRS. The rate of stroke or death was 10.3% at 5-years and 11.5% at 10-years after SRS. The only independent predictor of stroke or death in multivariate analysis was ‘increasing BAVM volume’. The risk of death or clinical impairment was 8.4% at 5-years and 12% at 10-years.


 The mean age and gender distribution are equivalent. The frequency of Spetzler-Martin subtypes were comparable, with the exception that grade IV patients encompassed 13% and grade V 0% of ARUBAs patients, while the composite of grade IV-V composed 19.5% of BAVMs of the report in question. The larger proportion of higher grade BAVMs (possibly including grade V lesions) in Pollock’s study may overestimate adverse event rates compared to what will be seen in ARUBA.

The authors provide us a solid and insightful analysis, which reflects the difficulties and uncertainties related to the treatment of BAVMs. As detailed, their rates of hemorrhagic stroke post SRS was comparable to the annual bleeding rate noted in the natural history of BAVM for the first five years, then decreasing below the risk for untreated BAVM as more patients achieve nidus obliteration. While we will have to wait for ARUBA to publish their results, this study can help us inform our patients about one current treatment for BAVMs.