Skip to content

Multimodal Approach to Unruptured Cerebral Arteriovenous Malformations Tops ARUBA Outcomes

Key findings

  • The ARUBA trial suggested that medical management is superior to intervention for unruptured cerebral arteriovenous malformations (cAVMs). However, this trial had significant limitations in both design and implementation
  • This retrospective study involved the largest multidisciplinary treatment cohort including 142 consecutive patients with cAVMs
  • The Department of Neurosurgery at Massachusetts General Hospital specializes in embolization, microsurgery and radiosurgery, the three treatments studied in the ARUBA trial
  • cAVM obliteration rates in the Mass General cohort were better compared to the ARUBA intervention arm (n=114) with far less disability and complications. While providing a cure for most patients, survival and disability rates were similar to the ARUBA medical management arm (n=109) after five years of follow-up

The optimal approach to the treatment of unruptured cerebral arteriovenous malformations (cAVMs) is hotly debated. The only randomized, controlled trial of this question, called the ARUBA trial, was halted in 2013 when an interim analysis suggested medical management alone was superior to intervention (embolization, microsurgery and/or stereotactic radiosurgery). However, the clinical experience at Massachusetts General Hospital showed high cure rates with low treatment complications that were similar to the natural history of patients.

There has been critical analysis of ARUBA, with concerns of insufficient enrollment, lack of standardization of treatment, underutilization of surgical resection, limited follow-up time and suboptimal choice of a primary endpoint (death or symptomatic stroke rather than permanent disability). In addition, numerous single modality treatment studies have documented lower disability and stroke rates than in the ARUBA intervention arm.

James D. Rabinov, MD, senior staff member in Interventional Neuroradiology and Endovascular Neurosurgery, and Bob S. Carter, MD, PhD, chief of the Department of Neurosurgery at Mass General, and colleagues recently reviewed data on ARUBA-eligible participants who were treated at Mass General. In the Journal of Neurosurgery, they report that with their multimodal approach, certain safety outcomes were significantly better than those reported for the ARUBA intervention arm (ARUBA Int) and similar to those in the ARUBA medical management arm (ARUBA Med) at five years of follow-up.

Study Design

The researchers retrospectively analyzed data on consecutive patients with a cAVM who were treated with surgery, radiosurgery and/or embolization between 2003 and 2015. They identified 142 patients who met ARUBA eligibility criteria: age 18 or older, unruptured cAVM diagnosed by neurovascular imaging, no previous treatment and a potential candidate for therapeutic cAVM eradication. The Mass General researchers further required a baseline score on the modified Rankin Scale (mRS) of 0 or 1 and evaluated five years of clinical follow-up.

Comparison of Outcomes to ARUBA

The researchers compared key outcomes in the Mass General cohort against those in the ARUBA intervention arm (n=114) and the ARUBA medical management arm (n=109):

  • Excellent clinical outcome (mRS score of 0 or 1) at five years: 87% (Mass General) vs. 60% (ARUBA Int) (P = .002) and 83% (ARUBA Med) (P = NS)
  • cAVM obliteration rate: 56% (Mass General) vs. 44% (ARUBA Int) (P = NS)
  • Symptomatic stroke (ischemic or hemorrhagic): 9% (Mass General) vs. 40% (ARUBA Int) (P = .0001) and 9% (ARUBA Med) (P = NS)
  • Significant disability (mRS score ≥2): 8% (Mass General) vs. 31% (ARUBA Int) (P = .004) and 13% (ARUBA Med) (P = NS)
  • Death: 2.1% (Mass General) vs. 3.8% (ARUBA Int) and 1.7% (ARUBA Med) (P = NS for both comparisons)

Stratification by Treatment Modality

All cAVMs treated with embolization and/or microsurgery were obliterated. Radiotherapy is often selected for inoperable cAVMs. By nature of its action an average of 2-4 years for follow up to assess efficacy.

Individual Treatment Modalities:

  • cAVM obliteration rate: Significantly higher in the Mass General cohort than in the ARUBA intervention arm when patients were treated with embolization alone (100% vs. 50%; P = .03) or radiosurgery alone (40% vs. 18%; P = .03)
  • Composite of death or symptomatic stroke (ARUBA primary endpoint): Significantly lower in the Mass General cohort than in the ARUBA intervention arm when patients were treated with radiosurgery alone (8% vs. 24%; P = .01) or with embolization plus surgery (5% vs. 43%; P = .03)

Cumulative Treatment Modalities:

  • cAVM obliteration rates: Significantly higher in the Mass General cohort than in the ARUBA intervention arm among patients who received any embolization (82% vs. 52%; P = .004) or any radiosurgery (39% vs. 21%; P = .03)
  • Death or symptomatic stroke: Significantly lower in the Mass General cohort than in the ARUBA intervention arm among patients who received any embolization (alone or with another intervention; 12% vs. 50%; P = .0002), any surgery (14% vs. 41%; P = .03) or any radiosurgery (11% vs. 37%; P = .002)

Clinical and Research Implications

Having one experienced neurosurgical team making recommendations and providing treatment provided excellent clinical results for patients with cAVMs. Superior results in the Mass General cAVM study compared to the ARUBA trial may be related to several factors. Tertiary referral centers such as Mass General, which have thorough expertise in all three modalities currently used to treat cAVM, may achieve better outcomes than those reported in ARUBA. Having one experienced neurosurgical team making recommendations according to a consistent algorithm and providing treatment is also important. A third organizational factor is the capacity to shift from one treatment modality to another after review of each stage is key to optimal care.

An important technical difference between the Mass General cohort and the ARUBA intervention arm is that endovascular treatment in ARUBA was intended to be primarily curative. For example, embolization is used principally as an adjunct to definitive surgical or radiosurgical resection compared to the ARUBA trial.

The length of follow-up also seems to have an important effect on outcomes. Surgical and procedure related complications occur at the time of AVM-targeted therapies. It is also known that efficacy for radiosurgery may take several years of follow up with some patients requiring a second treatment. Adding in the relatively low annual stroke rates in patients with unruptured cAVMs, it may take 5-10 years to see a benefit of an intervention.

For a future clinical trial, investigators should consider selecting patients who have risk factors for hemorrhage (e.g., deep venous drainage or deep location) and/or low risk of resection (Spetzler–Martin grade I or II) for an initial study or subgroup analysis. In such a patient population, intervention might demonstrate beneficial results at an earlier time.

Learn more about the Cerebrovascular Program at Mass General

Refer a patient to the Department of Neurosurgery


In this Q&A, Dr. Rost discusses her work on developing neuroimaging markers of cerebrovascular disease, stroke genetics and big data science for outcome prediction in patients with acute stroke.


MicroRNA miR-574-3p, which can be derived from whole blood, was associated with stroke in a community-based sample of more than 2,000 participants in the Framingham Heart Study's Offspring cohort.