Publications and Trials on Brain AVMs
Stapf et al. (2006) – Predictors of Hemorrhage in AVMs
Full Citation:
Stapf C, Mohr JP, Pile-Spellman J, et al. Predictors of hemorrhage in patients with brain arteriovenous malformation. Lancet Neurology. 2006;5(7):510–516. DOI: 10.1016/S1474-4422(06)70404-6
Stapf et al. published a notable prospective cohort study in Lancet Neurology in 2006 examining predictors of hemorrhage in patients with brain arteriovenous malformations (AVMs). The study emerged from the understanding that there was a dearth of high-quality prospective data from the natural history of unruptured AVMs, which is essential to determining when intervention might be appropriate. The authors hypothesized that particular anatomical and clinical features, in particular deep venous drainage, deep brain location, and the presence of prior hemorrhage would independently increase the risk of future AVM rupture in patients managed conservatively. A total of 722 patients with angiographically confirmed AVMs were enrolled at Columbia University Medical Center and followed prospectively, collecting data about demographic characteristics, the morphology of the AVM, and clinical presentation. The primary outcome of the study was symptomatic intracranial hemorrhage confirmed by imaging.
During follow-up, the investigators found an overall annual hemorrhage rate of 2.2% per year. The multivariate analysis identified three independent risk factors for future hemorrhage: by the presence of a prior ruptured AVM, deep venous drainage, and deep brain location. The size of the AVM and the age of the patients were not significant predictors of rupture risk. For patients with all three risk factors, the risk of hemorrhage was much higher, exceeding 10% per year. These data further supported the individualized risk stratification of AVMs in light of further management options and acknowledged that certain unruptured AVMs have an acceptably low annual bleeding risk and could be followed; whereas other AVMs, particularly those with prior bleeds or deep anatomical risk factors may require earlier intervention.
The study was one of the first to provide a strong, prospective estimate of the risk of AVM rupture in untreated patients and has been used since in informing practice for clinical decision making and in facilitating the design of future trials. It also served as validation for the inclusion of deep venous drainage and prior hemorrhage into AVM grading systems.
Kim et al. (2014) – Patient-Level Meta-analysis of Hemorrhage Predictors
Full Citation:
Kim H, Al-Shahi Salman R, McCulloch CE, Stapf C, Young WL. Untreated brain arteriovenous malformation: patient-level meta-analysis of hemorrhage predictors. Neurology. 2014 Aug 26;83(9):590–597. DOI: 10.1212/WNL.0000000000000688
To clarify such risk factors for hemorrhage, Kim and collaborators published a large patient-level meta-analysis in Neurology in 2014 designed to address some of the limitations that exist in prior single institution series. This meta-analysis reviewed data from up to 10 prospective cohorts throughout the world and, importantly, included data on AVMs both ruptured and unruptured. Kim et al. hypothesized that prior hemorrhage, deep venous drainage, and small AVM size would be independently associated with a higher risk of rupture. The final analysis included a total of 1,581 patients with unruptured or previously ruptured AVMs that were initially conservatively managed, making it one of the strongest datasets available on the natural history of AVMs.
After applying Cox proportional hazards models to examine the association between baseline features and hemorrhage risk we confirmed that a number of factors were significantly associated with rupture. The strongest predictor was history of a prior AVM hemorrhage, which increased the hazard of rupture to more than three times. Deep venous drainage was associated with a hazard greater than 1.7, and quite unexpectedly, AVMs smaller than 3 cm had associated risk hazard that was modestly high compared to larger AVMs. The models included factors such as deep brain location and age; however, these were not statistically significant predictors in all models. The overall annualized hemorrhage risk in this pooled cohort was consistent with prior reports, but did exhibit considerable variability based on the presence or absence of these key risk features.
The authors concluded that hemorrhage risk in untreated AVMs is heterogeneous and must be evaluated on a patient and vascular malformation-specific basis. Their conclusions further corroborated and add to prior work, such as the Stapf study, by confirming that prior hemorrhage and deep venous drainage is important in the assessment of risk. Clearly, the statistical power and population generalizability were strong from a large sample size and pooled prospective study design. The meta-analysis is still referenced in clinical practice guidelines and can serve as an evidence-based reference to determine if unruptured AVMs can be observed or should be treated. The study does not refer to treatment outcomes but establishes the use of natural history modeling as an important factor in the risk stratification of AVMs while establishing that the authors’ variables are warranted for inclusion in decision-making algorithms.
Spetzler & Martin (1986) – Surgical Grading System for AVMs
Full Citation:
Spetzler RF, Martin NA. A proposed grading system for arteriovenous malformations. Journal of Neurosurgery. 1986 Oct;65(4):476–483. DOI: 10.3171/jns.1986.65.4.0476
In 1986, Spetzler and Martin, presented a comprehensive surgical grading scale for brain arteriovenous malformations that aimed to estimate operative risk. The authors were motivated to conduct the study due to no consistent approach to guide surgical decisions pertaining to AVM treatment. The authors identified three anatomical characteristics as potentially useful in systematically assessing the risk of surgical morbidity and mortality. Therefore, in order to test the usefulness of these characteristics, the authors retrospectively reviewed 100 adults with a brain AVM, who underwent microsurgical exploration at the Barrow Neurological Institute. Each AVM was scored on three attributes: size, eloquence of the adjacent brain, and nature of the venous drainage. The scores for each of these characteristics were then totalled to determine the Spetzler-Martin grades I through V.
The surgical results were stratified by grade, and the data showed a strong association between higher Spetzler-Martin grade and increased complication rates. Specifically, the authors reported that:
- Grade I (n=4): 0% surgical mortality; 0% permanent morbidity
- Grade II (n=39): 0% mortality; 0% permanent morbidity
- Grade III (n=34): 0% mortality; 11.8% permanent morbidity
- Grade IV (n=20): 15% mortality; 20% permanent morbidity
- Grade V (n=3): 33% mortality; 100% permanent morbidity
These findings emphasized that AVMs of Grade I and II were very compatible with surgical treatment and had excellent prognosis while Grades IV and V have an increased risk of surgical mortality and morbidity, and were usually poor surgical candidates. Grade III appeared to be borderline and required more individual consideration.
The authors concluded that this three-variable grading scale provided a clinically relevant, reproducible method of stratifying surgical risk of patients with AVM. The Spetzler-Martin scale gained a wide audience after its publication and currently represents the most common way of grading surgical risk in the management of AVM. The Spetzler-Martin scale does not account for patient-specific and AVM factors such as age or rupture status, and its ease of use and predictive ability has made it a cornerstone of clinical decision making and treatment algorithms, especially regarding surgical management of AVM versus radiosurgery or embolization in a multidisciplinary approach to AVM management.
Lawton et al. (2010) – Modified Spetzler-Martin Supplementary Scale
A landmark study published by Lawton et al. in Neurosurgery in 2010 sought to enhance surgical decision making for patients with brain arteriovenous malformations undergoing resection. While the Spetzler-Martin scale served as a risk scoring system for demographic and anatomical predictor variables, it did not include important patient-specific variables that could also influence surgical outcomes. The authors hypothesized that if they included more clinical variables, then surgical outcomes could be more accurately predicted. To assess if this would indeed improve predictive assessment they retrospectively reviewed a total of 300 consecutive patients who had undergone surgical AVM resection at the University of California, San Francisco. Patients received a score using a newly proposed supplementary scoring scale.
- Age: less than 20 years = 1, 20-40 years = 2, greater than 40 years = 3;
- Rupture: ruptured = 0, not ruptured = 1;
- Compactness: compact = 0, diffuse = 1.
- The total supplementary score (0 to 5) was then added to the SM grade (1 – 5) to generate a cumulative risk score (1 – 10).
Key statistics showed this extra score was predictive of outcome. Specifically the extra score independently predicted surgical morbidity, with each extra point increasing the risk of poor outcome. Stratified outcomes were impressive: 0-1 score had very low permanent neurological deficits (3-5%), 2-3 score had moderate morbidity (10-20%), 4-5 score had much higher risk with permanent morbidity 30-50%. When the extra score was added to the SM grade the predictive accuracy of adverse outcomes improved.
The authors found this supplementary grading system to be a game changer for surgical risk stratification for AVM patients. By including clinical factors like age, rupture history and nidus compactness the scale allowed for more precise and individualized counseling. The Lawton-Young Supplementary Scale has since become an essential tool in AVM management especially for borderline or intermediate grade (Grade III) lesions. Although retrospective, the statistical validation and clinical practicality of this scale has improved neurosurgical decision making and has been widely adopted and included in clinical guidelines.
Morgan et al. (2017) – Meta-analysis of Modern AVM Surgery Outcomes
Full Citation:
Morgan MK, Davidson AS, Assaad NNA, Stoodley MA. Critical review of brain AVM surgery, surgical results, and natural history in 2017. World Neurosurgery. 2017 Oct;106:948–959. DOI: 10.1016/j.wneu.2017.07.093
Morgan and colleagues did a meta-analysis published in World Neurosurgery in 2017 to critically evaluate current surgical outcomes and refine our understanding of surgical efficacy and risk of brain AVMs. Recognizing advances in microsurgical techniques, patient selection and imaging modalities the authors hypothesized that modern surgery especially for low grade AVMs (Spetzler-Martin Grade I-II) would have high obliteration rates with minimal complications. They reviewed 16 high quality surgical series published between 2000 and 2017 with over 3,000 patients who underwent microsurgical AVM resection.
The statistical analysis of the combined data showed a clear and significant relationship between Spetzler-Martin grade and outcome. For Spetzler-Martin Grades I–II AVMs, microsurgical resection had extremely high complete obliteration rates, usually above 90–95% with less than 5% permanent morbidity. Grade I AVMs had nearly 100% resection (~98% obliteration) and Grade II AVMs had high success (~94% obliteration). Grade III AVMs had moderate risk, with 90% obliteration but 10–15% permanent morbidity. For higher grade AVMs (Grades IV–V) the risk-benefit ratio was much less favourable, with lower obliteration rates (~70–80%) and 30–50% permanent morbidity and mortality (often above 5%).
According to Morgan et al., microsurgery proves to be both effective and safe for low-grade AVMs, making it a strong recommendation for treating Spetzler-Martin Grades I–II lesions. They did, however, urge caution when it comes to higher-grade AVMs, where a more comprehensive approach such as embolization or radiosurgery might be necessary due to the higher risks associated with surgery. Their findings stressed the importance of selecting patients carefully and reaffirmed the value of the Spetzler-Martin grading system in making clinical decisions. Despite some limitations from retrospective analyses and the diversity of the studies reviewed, this meta-analysis delivered solid statistical support for the outcomes of contemporary AVM surgeries, significantly shaping clinical practices and patient guidance.
Mohr et al. (2014) – ARUBA Trial
Full Citation:
Mohr JP, Parides MK, Stapf C, et al. Medical management with or without interventional therapy for unruptured brain arteriovenous malformations (ARUBA): a multicentre, non-blinded, randomised trial. The Lancet. 2014 Feb 15;383(9917):614–621. DOI: 10.1016/S0140-6736(13)62302-8
The ARUBA trial, which stands for A Randomized Trial of Unruptured Brain Arteriovenous Malformations, was published in The Lancet back in 2014. The goal was to figure out whether treating unruptured brain AVMs with medication alone or using more invasive methods like surgery or embolization would lead to better health outcomes. The researchers, led by Mohr and his team, believed that medical management could actually prevent more strokes and deaths compared to the interventions, which come with their own risks.
A total of 223 adult patients with unruptured AVMs were enrolled and randomized into 109 patients to conservative medical management alone (observation and medical therapy) and 114 patients to interventional treatment. The composite primary endpoint was the occurrence of symptomatic stroke or death from any cause, and patients were followed for a median of about 33 months. The trial was terminated early based on a large difference in primary endpoint event rate. The medical management arm had a significantly lower event rate (10.1% at 33 months) vs the intervention arm, which had a significantly higher event rate (30.7%). This translated into a hazard ratio (HR) of 0.27 favoring conservative management (95% CI: 0.14–0.54, p < 0.001).
The ARUBA trial showed that medical management seemed better in the short term, but it also got some criticism. One main point raised was that the average follow-up of just under three years wasn’t long enough to really understand the long-term risk of AVM rupture, especially since these AVMs usually impact younger people. Another issue was that the treatment strategies in the intervention group varied a lot, which could have skewed the results. Plus, they expected to have 800 patients but only managed to randomize 223, which brings up concerns about selection bias and how well the results can apply to a broader group.
Despite these criticisms, ARUBA continues to be relevant, reminding us that the short-term risks of treating unruptured AVMs might outweigh any long-term benefits we don’t know about yet. Throughout the study, there was a clear understanding of the importance of careful, tailored evaluations and treatment plans for patients. It was especially noted that if a patient has an asymptomatic AVM, it’s wise to approach any aggressive treatment options with caution.
Starke et al. (2017) – Radiosurgery Outcomes in AVMs
The ARUBA trial showed that treating unruptured arteriovenous malformations with medical management offered better short-term results, but it raised some important questions about how it was set up. One of the main critiques was that the follow-up time of just under three years might not be enough to really understand the long-term risk of AVM ruptures, especially since these issues often hit younger people. Also, the treatment methods varied a lot due to outside specialists being involved, which might have affected the results. Plus, they only managed to enroll 223 patients instead of the planned 800, raising concerns about whether the findings really apply to a broader patient group.
Despite these concerns, ARUBA serves as a reminder that short-term risks of treating unruptured AVMs can outweigh potential long-term benefits. Throughout the trial, there was a clear understanding that each patient needs careful, personalized evaluation and management, especially for those with asymptomatic AVMs.
In 2017, Starke and his team shared a study in the Journal of Neurosurgery that looked at long-term results of using stereotactic radiosurgery for AVMs. With radiosurgery becoming a popular noninvasive option, especially for smaller or tricky-to-reach AVMs, they believed it could be a safe and effective treatment with high success rates and limited long-term issues. Their research combined data from eight centers around the world, covering 2,236 patients treated between 1987 and 2014, with an average follow-up time of about seven years.
Most patients had small-to-medium sized AVMs, with an average volume of around 2.4 cm³. The results were promising, showing that 64.7% of patients had complete AVM closure after five years and this number rose to about 75% after ten years. After the AVMs were confirmed to be closed, the risk of bleeding dropped significantly to under 1% per year. However, in the period before complete closure, the bleeding risk was still around 1.4%, indicating that radiosurgery has short-term risks compared to more immediate results you get from surgery. About 7.4% of patients experienced serious complications from radiation, with around 3-5% developing symptomatic radiation necrosis.
The authors found that SRS is a good long-term treatment option, especially for small AVMs in tricky areas of the brain where surgery can be risky. They stressed that patients need to be aware that it takes time to see results and that there’s still a chance of bleeding during the waiting period. This big study backed up SRS as a safe and lasting treatment when chosen wisely based on the patient’s situation and the details of the AVM. This paper has made a real impact on how doctors treat AVMs today, especially for the right patients.
Saatci et al. (2011) – Onyx Embolization Outcomes
In 2011, Saatci and his team published a study in the Journal of Neurosurgery that looked at the long-term effects of using a liquid embolic agent called Onyx for treating brain arteriovenous malformations. They believed that a specific injection method with Onyx would help it reach deeper into the AVM, leading to more successful treatment outcomes while keeping complications low.
The research included 350 patients who underwent embolization over a decade at a busy neurointerventional center. The procedures were done in stages, with a focus on fully sealing the AVM.
The results showed that 179 out of the 350 patients, or about 51.1%, had complete AVM obliteration from the endovascular treatment alone. Success rates varied based on the Spetzler-Martin grade of the AVMs: the lowest-grade lesions had cure rates of 70-80%, while higher-grade lesions had lower success at around 40%. Complications from embolization were rare, with a permanent morbidity rate of only 1.7% and a procedure-related mortality rate of about 1.1%. During follow-ups, only around 1.4% of patients experienced symptomatic bleeding. Importantly, those who achieved an angiographic cure had a very low risk of bleeding after treatment, less than 0.5% each year, indicating lasting safety.
The authors found that their careful method of injecting Onyx into blood vessels led to much better results in fully blocking off certain types of vascular issues compared to standard methods. This was especially the case for less severe conditions, and they noted that the risks of lasting complications were low. They pointed out that when done by skilled professionals, this type of treatment could be a go-to option for specific cases, particularly those that are easier to treat with less invasive techniques. Still, they mentioned that for larger or more complicated issues, a mix of treatments might be necessary since embolization alone might not be enough. This solid study really showed how Onyx embolization is becoming an important tool for treating brain vascular problems today.
Interested in learning more about our practice?
