Introduction

Arteriovenous malformations (AVMs) are detected at a rate of approximately 1 per 100,000 person-years. AVM-related hemorrhage as identified in the prospective trials of the New York Islands AVM hemorrhage study group and the Northern Manhattan stroke group has an incidence of approximately 0.42 and 0.55 per 100,000 person-years, respectively, and accounts for nearly 40% of intracerebral hemorrhage (ICH) in patients aged between 15 and 45 years. AVMs can present throughout the central nervous system often with variable vascular and surrounding anatomy and thus provide a formidable challenge to the treating neurosurgeon. While operative intervention has been the mainstay of treatment for low-grade AVM (grades I and II), other treatment modalities such as stereotactic radiosurgery (SRS) and endovascular embolization are used judiciously in appropriate cases.

Major controversies in decision making addressed in this chapter include:

1. 
   

   Whether or not treatment is indicated.

   
   
2. 
   

   The right timing of intervention.

   
   
3. 
   

   The type of treatment(s).

   
   
4. 
   

   The treatment for ruptured grade I and II AVMs.

Whether to Treat

The decision as to whether a patient requires treatment depends on balancing the risk of treatment with the risk inherent in the natural history of the pathology. This is further complicated in AVM treatment as often the risk of treatment will be taken up-front, while the risk of conservative treatment comes at a later time. Thus, any decision needs to be patient centered and will require detailed discussions between the treating physician and the patient.

The overall rupture rate for any AVM is approximately 3% per year. This can be further stratified according to whether the AVM has previously ruptured ( 1 , 2 in algorithm ). For unruptured AVMs, multiple studies, including the medical management with or without interventional therapy for unruptured brain AVMs (ARUBA) trial, have found that the risk of rupture is slightly less, at approximately 2% per year (approximately 10–11% over 5 years; 1 , 2 in algorithm ). Risk factors, which have been associated with rupture risk, include a deep location of the AVM, younger age, exclusive deep venous drainage, associated aneurysms, venous stenosis, diffuse AVM anatomy, and high arterial feeder pressure. Other risk factors that have been less well defined include sex, ethnicity, vertebrobasilar supply, and hypertension ( 2 in algorithm ). Whether pregnancy increases the risk of AVM rupture is a matter of ongoing debate.

For AVMs that have already ruptured, the risk is quite different. One of the major risk factors for hemorrhage is previous hemorrhage. According to a number of retrospective studies, the risk of a subsequent hemorrhage may be as high as 7 to 10% per year. It appears that the risk is highest in the initial 2 years following a hemorrhage and then it subsequently declines over the subsequent 5 years ( 1 in algorithm ). However, the risk of further rupture remains greater than that of an unruptured AVM for the remainder of the patient′s life.

While the risk of AVM rupture appears to be relatively low when compared to the risk of aneurysm rupture, the consequence can be devastating. Further, AVM hemorrhages tend to occur in younger patients and can have lasting complications. In a study of 115 patients suffering AVM hemorrhage, 84% of the patients had a modified Rankin Scale (mRS) score of 1. The risk of rerupture varies between 6 and 18%. Mortality from rehemorrhage is greater with some reports stating that up to half of these patients die ( 1 , 2 in algorithm ). Other risk factors to consider include the presence of seizures (which have an associated morbidity and mortality) and neurological deficits, which can occur secondary to vascular steal phenomena or mass effect, although these are quite rare.

The cumulative survival rates of AVM patients when compared to the general population (1.0) after 30 years of follow-up have been 0.49 in those with untreated AVMs and 0.87 in those with obliterated AVMs, suggesting treatment of the AVM with obliteration improves survival in the long term ( 6, 8, 9, 10 in algorithm ). The ARUBA trial, a multicenter, nonblinded, randomized trial, which compared medical management to a variety of interventions, found that the risk of using standard medical management (essentially natural history) was significantly less than that associated with intervention. However, this study was limited in a variety of ways, particularly by the short-term follow-up, which biases the interventional cohort who takes their risks up-front and the very limited number of cases treated with surgery alone (< 5 cases). Most cases were treated with radiosurgery or endovascular treatment alone or with a combination of both and were not treated to complete obliteration.

We have previously discussed the factors that increase the risk of morbidity and mortality from surgical intervention, namely, size of the AVM, pattern of venous drainage, eloquence of the area surrounding the AVM, age of the patient, whether prior rupture has occurred, diffuseness of the AVM, and whether there is a perforating arterial supply. In general, morbidity and mortality for treated grade I and II AVMs is very low (<1% of patient will have a major deficit).

In summary, elective surgery is appropriate for grade I and II (class A) AVMs if the surgical anatomy and eloquence of the surrounding brain are favorable given the 2% annual risk of rupture ( 6, 8, 9, 10 in algorithm ). Given the early results of the ARUBA trial, it is unclear what the role of endovascular or radiosurgery has in this cohort of patients, particularly given that surgical outcomes suggest mortality and morbidity to be less than 1%.

When to Treat

Once a decision to treat has been made, the timing of treatment becomes the next important consideration. For unruptured AVMs, the 2% per year rupture rate generally means that a thorough evaluation and workup can be conducted, followed by a detailed discussion with the patient. Generally, given the low morbidity and mortality for surgery for grade I and II AVMs, surgery should proceed in a timely manner to maximize the reduction in AVM rupture risk ( 9 in algorithm ).

For ruptured AVMs, many considerations influence the choice and timing of treatment. For obtunded patients with raised intracranial pressure (ICP), intervention may include cerebrospinal fluid (CSF) diversion, decompressive craniectomy, or urgent hematoma evacuation ( 3, 5, 6 in algorithm ). The hematoma should be evacuated to identify the potential source of hemorrhage. If there is an obvious intranidal aneurysm or other likely source of hemorrhage, this can be targeted with initial endovascular or surgical treatment to prevent early rehemorrhage.

This may then be followed by definitive AVM treatment if the patient survives. While the timing of this treatment is controversial, many experts would suggest treatment 2 to 4 weeks following the ictus, which allows the hematoma to liquefy, thus providing an additional nonanatomical space to use for access during resection ( 8 in algorithm ).

For neurologically intact patients suffering AVM hemorrhage, a rapid but thorough workup is indicated. Again, timely surgical management in patients with appropriate anatomy would be recommended. Given the increased risk of rehemorrhage, workup and subsequent treatment should be expedited.

Classification

While a number of grading systems exist and have been validated in the neurosurgical literature, the most commonly used system is the Spetzler–Martin grading system (▶ Fig. 53.1 ). The system assesses three major anatomical factors, namely, the size of the nidus of the AVM (▶ Figs. 53.1a, d ), the pattern of venous drainage (superficial or deep; ▶ Fig. 53.1c ), and the neurological eloquence of the brain regions adjacent to the AVM (▶ Fig. 53.1b ; ▶ Tables 53.1 , 53.2 ). Using this system, each AVM can be given a score between 1 and 5 that subsequently produces a five-tiered system that correlates the likely neurosurgical morbidity and mortality from operative intervention. Low-grade (grades I and II) AVMs have a low operative morbidity and mortality (<1% of patients will suffer a major neurological deficit or death). Grade III lesions are of intermediate risk, with risks being higher for larger AVMs in eloquent regions of the brain. Grade IV and V lesion have a high risk of postoperative neurological deficits. Further refinements have been made to this grading system. By combining grade I and II AVMs and grade IV and V AVMs, a three-tiered classification scale has been created, providing a more robust system. Similarly, a supplementary classification system that incorporates elements of age, previous rupture, diffuseness, and perforating arterial supply in addition to size, venous drainage, and eloquence has been used to further stratify patient risk.

A grading system for radiosurgery treatment has also evolved and is based on the volume of the AVM, the age of the patient, and location of the AVM. As the score increases, the likelihood of an excellent outcome (obliteration without neurological injury) decreases. Throughout this chapter, we will be discussing grade I and II AVMs as graded by the five-tiered Spetzler–Martin grading system (Spetzler–Ponce class A AVMs). Approximately 50 to 60% of all AVMs are categorized as a grade I or II.