Update on the Statement on Brain AVMs: Despite New Data, Questions Still Unanswered

First, in any discussion of brain arteriovenous malformations (bAVMs), one would be remiss not to acknowledge our departed colleague, Dr. William Young, who contributed so greatly to our knowledge of this disease.

The updated AHA/ASA Scientific Statement on the Management of Brain Arteriovenous Malformations aims to review the current data and make recommendations on the diagnosis and management of ruptured and unruptured bAVMs¹. While a comprehensive review of the current literature, the statement is limited, as the writing group would likely freely admit, in making meaningful management recommendations because of the paucity of high quality evidence. This is in part due to the rarity of this disease. Brain AVMs are rare, occurring in 10-18 per 100,000 adults,^2,3 with new bAVM detection occurring at a rate of 1.3 per 100,000 person-years^4-6.

This update to the previous AHA/ASA Scientific Statement on Intracranial AVMs (2001)⁷ is undoubtedly prompted by new data from a randomized controlled trial of medical management versus interventional therapy for unruptured bAVMs. A Randomized Trial of Unruptured Brain AVMs (ARUBA) randomized 226 adults with unruptured bAVMs from 39 worldwide clinical sites over a 6-year enrollment period to medical management alone or medical management with interventional therapy⁸. The primary endpoint of death or symptomatic stroke occurred in 11 of 109 patients in the medical arm (10.1%) and 35 of 114 patients in the interventional arm (30.7%)(HR 0.27, 95% CI 0.14-0.54).

As the writing group points out in this scientific statement, there are limitations of the ARUBA study:

1. Short follow-up period: The average follow-up in ARUBA was 33 months which is inherently biased towards medical management because the risk of hemorrhage and disability with unruptured bAVMs may be relatively low, but lifelong, and thus, the benefit of interventional therapy would be realized over a longer term.
2. Treatment modalities: The treatments rendered in the interventional therapy arm were biased towards stereotactic radiosurgery (n=31) which has a well-known latency interval to therapeutic effect, and embolization alone (n=30) which has lower curative success and higher complication rates than surgery (surgery alone, n=5; embolization followed by surgery, n=12).

The writing group did not discuss these additional limitations of ARUBA which deserve mention:

1. Selection bias: Of the 726 eligible screened patients, 226 were randomized, 323 refused enrollment, and 177 were managed outside of the randomization process. Unwillingness to participate in randomization can be influenced by concerning characteristics of the bAVMs.
2. Distribution of Spetzler-Martin grades: Spetzler-Martin (SM) grades in ARUBA were SM-I (29%), SM-II (32%), SM-III (28%), SM-IV (10%). As the writing group points out in their discussion of SM grades, surgical resection is best suited for SM-I and SM-II bAVMs. SM-IV bAVMs are rarely surgically resected because of the known morbidity and mortality risks, and SM-III bAVMs represent a heterogeneous group with modified SM-III (medium size, non-eloquent, deep venous drainage) and modified SM-III+ (medium size, eloquent, superficial venous drainage) bAVMs, having worse surgical outcomes similar to high-grade bAVMs9. Indeed, the SM-I bAVM patients in ARUBA fared better in the interventional arm (6.3% with endpoint) than the medical management arm (12.1% with endpoint), though not statistically significant due to lack of power with the small sample size.

What we can learn from the ARUBA study is that the spontaneous bAVM rupture rate in the medical management arm was 2.2% per year which is consistent with previous unruptured bAVM natural history studies^10-14.

This scientific statement update reiterates that for unruptured bAVMs, they recommend informing patients of the natural history risk of hemorrhage and seizure, discussing treatment options weighing risks of treatment and life-expectancy, and using the Spetzler-Martin grading scale to predict risk of surgical resection.

Computed tomographic angiography (CTA), magnetic resonance angiography (MRA), and digital subtraction angiography (DSA) are recommended in patients with nontraumatic intracranial hemorrhage to evaluate for potential underlying bAVMs.

For ruptured bAVMs, the AHA/ASA Guidelines for the Management of Spontaneous Intracerebral Hemorrhage¹⁵ recommends initial management of the hemorrhage and considering treatment options for bAVMs by weighing the relative risks and benefits.

In summary, the writing group should be congratulated on a comprehensive review of the current literature and data on bAVMs. There have been significant advances in knowledge about the biology and genetics of bAVMs, better imaging capabilities, new endovascular agents, and added cumulative experience with surgery and stereotactic radiosurgery. However, the management recommendations are largely unchanged from the previous AHA/ASA scientific statement 16 years ago. This reflects the lack of high quality evidence, despite the addition of new data from a randomized controlled trial, and most questions are still left unanswered in guiding clinical management of bAVM patients.