♦ Epidemiology

Intracranial AVMs are thought to be congenital lesions with one-tenth the incidence rate of intracranial aneurysms. The true incidence is difficult to assess because the coding system of the International Classification of Diseases (ninth revision) does not classify iAVMs separately, often grouping these lesions with several other kinds of intracranial vascular malformations, including but not limited to cavernous malformations and unruptured aneurysms. Furthermore, initial admissions for intracerebral hemorrhage or seizure can often go unrecognized as being caused by the presence of an iAVM. Several autopsy series have reported prevalence data ranging from 5 to 613 AVM cases per 100,000.3 ^, 7 ^, 8 Retrospective studies have shown the iAVM incidence to be 0.51 to 5 per 100,000.9 ^– 11 Prospective studies, such as the New York Islands Arteriovenous Malformation Hemorrhage Study and the Manhattan Stroke Study, reported that the average annual iAVM detection rate was 1.34 per 100,000 personyears (95% confidence interval [CI], 1.18–1.49) and 0.55 per 100,000 person-years (95% CI, 0.11–1.61), respectively.12 ^, 13 Despite relative agreement between retrospective and prospective studies on the incidence and prevalence of iAVMs in the general population, the actual patient burden of iAVMs may never be known due to the underreported incidence of asymptomatic iAVMs or hidden iAVMs in the setting of fatal or very large hemorrhages without angiographic evaluation. Although relatively rare, these lesions are an important and preventable cause of hemorrhagic stroke, death, and disability in a relatively young patient cohort.

♦ Natural History

The natural history of iAVMs remains a controversial topic. Intracranial AVMs are a heterogeneous group of vascular malformations that elicit highly variable characteristic behavior and that depend on the angioarchitectural features of the lesions themselves, such as large parent artery feeders, associated venous stenosis, nidal or pre-nidal associated aneurysms, and so on. Furthermore, iAVMs present clinically in a variety of ways, most commonly hemorrhage, seizure, headache, or focal neurologic deficits, or can be found incidentally. Hemorrhage rates for iAVMs range from 2 to 4%1 ^, 14; however, it can increase in certain specific circumstances, such as pregnancy, previous hemorrhage, progressive venopathy, or intranidal aneurysms that are arterial or venous.

In the New York Island study, Stapf and colleagues13 ^, 15 reported that hemorrhagic presentation, older age, deep location, and exclusive deep drainage were independent predictors of hemorrhage. The highest risk population had a hemorrhage rate up to 34% within a mean follow-up of only a few months. Young patients with superficial iAVMs and superficial venous drainage were noted to have lower hemorrhagic risk. These observations suggested that the hemorrhagic risk is a dynamic characteristic of iAVMs and likely depends on both local and systemic hemodynamic changes. In contrast, Hernesniemi et al1 recently reported that the highest risk of hemorrhage was found in young patients during the first 2 years after initial diagnosis. The annual hemorrhage risk is 4.6%, during those 2 years and decreases to 1.6% in subsequent years. Both studies, however, did agree that previous hemorrhage, deep location, and exclusively deep venous drainage were independently associated with a higher rupture risk. Timely management was advised for lesions that fit this description. However, the treatment of the more benign iAVMs (e.g., located in non-eloquent or superficial regions, exclusively superficial venous drainage, and no prior history of hemorrhage) is particularly controversial due to the lack of knowledge on the natural history of these lesions. For a more complete understanding of the natural progression of iAVMs, future studies will require long-term followup (>15 years) because the incidence and annual hemorrhage risk is low. Studies involving follow-up periods of <5 years are unlikely to provide the information necessary to correctly elucidate the true natural history of iAVMs. In addition, radiographic subtleties such as nidal aneurysms, venous stenosis, and angiomatous versus embryonal morphology should be carefully studied. At this time, the discussion of the natural history of iAVMs versus the surgical risks involved can be difficult to compare and contrast based solely on the literature and often becomes reliant on the experience and intuition of the treating physicians.

♦ Patient and Lesion Selection

The decision to treat an intracranial AVM must involve balanced considerations of both lesional and patient characteristics ( Table 23.1 ). When to treat an intracranial AVM is often a more difficult question than how to treat, except when patients present with hemorrhage, unrelenting seizure, or progressive neurologic deficits. Patient-specific factors such as age, comorbidities, weight, occupation, psychological burden, and risk aversion must be completely evaluated. Lesion-specific factors such as size, functional eloquence of location, venous stenosis, associated aneurysms, corridors of access, angiomatous changes, and dysplastic arterial supply must be carefully assessed ( Fig. 23.1 ). Asymptomatic iAVMs that present with angioarchitectural features associated with higher hemorrhagic risks, such as venous stenosis, associated aneurysms, and deep drainage, are recommended for evaluation and possible intervention. Subsequently, a risk-benefit paradigm can be assembled predicated on a best-guess analysis of natural history risk if left untreated. Intracranial AVMs that are cortically based, that involve only superficial drainage, and that do not involve eloquent cortex have a low surgical morbidity. However, the superficial drainage may be draped over the nidus and pose increased surgical risks. In the same fashion, a deep AVM may pose a higher surgical risk, but if associated with a previous hemorrhage, the hematoma may provide a surgical corridor that decreases associated surgical morbidity. Deep venous drainage, which has traditionally been associated with increased surgical morbidity, may in fact be advantageous such that these veins do not impede surgical resection of the nidus. AVMs that are anatomically within an eloquent area, as shown by functional imaging, will likely pose significant surgical risk and are often relegated to alternative treatments. In an earlier report, Pollock et al classified AVM patients into four groups, based on rupture risk:

1. 
   

   Low-risk AVMs: no history of a prior bleed, >1 draining vein, and a compact nidus

   
   
2. 
   

   Intermediate-low-risk AVMs: no history of a prior bleed, 1 draining vein, and/or a diffuse nidus

   
   
3. 
   

   Intermediate-high-risk AVMs: history of a prior bleed, >1 draining vein, and a compact nidus

   
   
4. 
   

   High-risk AVMs: history of a prior bleed, 1 draining vein, and/or a diffuse nidus

The low-risk group had an annual rupture risk of approximately 1.3% and a 40-year risk of 33%, whereas the high-risk group had an approximately 9% annual risk and a 40-year risk of 98%. The authors used a predictive formula that assumed that each hemorrhage is an independent event and that there are no deaths from AVM hemorrhage:

However, surgical morbidity and mortality weighed against the natural history of the disease are not the only factors to be taken into account when making treatment decisions. The patient’s wishes and expectations must also be considered. Despite the fact that the annual hemorrhage risk of most iAVMs is small, it is the lifetime accumulation of risk that should be discussed and compared with an upfront procedure risk. Patients often have difficulty comprehending lifetime risk potentials regarding death and disability, but the concepts must be communicated.

♦ Surgically Important Angioarchitectural Characteristics

Detailed assessment of the iAVM angioarchitecture is critical to the success of surgical management. Preoperative imaging provides the angioarchitectural features, such as arterial feeder localization, nidal morphology, and draining venous patterns, which the surgeon will be required to know and anticipate at every point during the surgical resection ( Fig. 23.1 ). The iAVM nidus can be classified into compact, diffuse, or a combination, usually characterized by separated pedicles, occurring naturally, from previous treatment (endovascular or radiosurgery), or from previous hemorrhage. A compact nidus is desirable from a surgical perspective and usually allows a more manageable resection because there is little to no brain parenchyma involved within the nidus, and a perinidal subarachnoid dissection plane can be more easily established. Arteriovenous malformations with a diffuse nidus are inherently more difficult to manage surgically, particularly if near eloquent tissue because of the indistinct nature of the nidus and the intervening brain parenchyma between tufts of the nidus. The characterization of the arterial supply to an iAVM nidus is also extremely important, as it will dictate not only the nidus location, but also what to anticipate during surgical resection. An attempt must be made to preoperatively identify the presence of en passage vessels because inadvertent ligation of these vessels can put distal tissue at risk. Parasylvian iAVMs, in particular, routinely have middle cerebral artery (MCA) trunks that feed the malformation but continue on to supply other vital parts of the brain. Further, deep-feeding vessels can be anticipated and better controlled when the iAVM nidus itself or the apex is located near an ependymal surface. These feeders must be anticipated to avoid significant bleeding, which can be difficult to control and can quickly fill the ventricular system. Lastly, extracranial arterial feeders should be identified because these feeding vessels can usually be preoperatively embolized. Extracranial feeding arteries with transosseous components can complicate the craniotomy and, if not anticipated, can lead to serious epidural bleeding during dissection and elevation of the bone flap.

Identification and characterization of the venous angioarchitecture of the iAVM is as important as or even more important than the identification and characterization of the arterial system. Although superficial draining veins usually portend a lower hemorrhage risk profile, they can often make surgical resection more difficult, especially if they are draped across the surgical field. Retraction of these superficial veins and repetitive coagulation near them can increase the turgor of the iAVM, in spite of progressive deafferentation, thereby increasing the risk of intraoperative hemorrhage. Deep draining veins, although thought to increase the hemorrhage risk, actually provide a surgical advantage, as they will not be in the surgical field when dissecting the iAVM nidus. If superficial in origin, the main draining vein should be identified and protected throughout the surgical resection of the nidus, as it will be the last vessel ligated. However, depending on the pedicle construction of the iAVM nidus, there may be more than one main draining vein. These draining veins should also be maintained until all of the feeding arteries to that particular pedicle are ligated and cut. Premature ligation of important draining veins puts the pedicle at higher risk of congestion, leading to potential intraoperative rupture. As the procedure progresses, it is sometimes necessary to ligate a vein even while arterial supply persists. Placing a temporary clip on the vein and observing the iAVM and palpating it for turgor assessment can give vital information as to the safety of dividing a particular draining vein.

♦ Preoperative Embolization

Endovascular therapy of iAVMs offers great assistance to the surgical treatment.16 ^– 26 Strategies for endovascular therapy include the embolization of flow-related and peri-nidal aneurysms, progressive flow reduction to reduce the hemodynamic impact of “sudden” iAVM obliteration, and selective obliteration of deep feeding. Rarely, iAVMs may be cured by endovascular therapy alone, but partially treated iAVMs are thought to maintain similar hemorrhagic risks as untreated lesions.

Intracranial aneurysms are associated with iAVMs in approximately 30% of cases, half of which are on arterial feeders and half are located in the nidus itself.5 ^, 27 ^– 31 The timing of treatment of these aneurysms has been often debated. At our institution, we follow a few basic rules to determine the treatment of aneurysms associated with AVMs. If the aneurysm is associated with a pedicle of the AVM or located on the proximal circle of Willis and thought to be the cause of hemorrhage, then that aneurysm is treated acutely. If the aneurysm is located within the AVM nidus and is thought to have ruptured, then the aneurysm is treated early, if feasible. If the aneurysm is located on the proximal feeding arteries to the nidus and treatment of the AVM may put the aneurysm at risk of rupture by increasing outflow resistance and transmural gradient, then the aneurysm is treated via endovascular or microsurgical means, if feasible prior to AVM treatment. Finally, if inaccessible aneurysms exist that are thought to be flow-related and small, usually on the order of <3 mm, they are usually left alone and typically resolve on their own once the AVM is treated.

Careful study of the lesional morphology and the specific surgical problems that will be encountered help set the stage for developing an endovascular strategy for each unique patient. The progressive development of endovascular techniques and embolic agents has transformed the surgical management of these lesions. Each trip to the endovascular suite, however, may carry a 5 to 8% risk of ischemic or hemorrhagic complications (likely lower in careful hands). Large, high-flow iAVMs may be selected for slow, repetitive embolization procedures with 1 to 3 weeks between each treatment to allow for incremental flow redistribution to the surrounding brain tissue. For smaller lesions, embolization may target medial and lateral posterior choroidal feeders, for example, as these are the deepest areas to access during resection. We prefer aggressive embolization in parasylvian iAVMs as it allows for intraoperative road mapping and preservation of en passage vessels. The value of embolization of eloquent iAVM margins is that the surgical planes can be developed with more precision. External feeders are embolized to simplify the craniotomy, as mentioned above. In many small cortical iAVMs with superficial arterial supply, the risk of embolization may not be justified unless endovascular cure is the goal. The bottom line is that the treating surgeon should serve as the “captain of the ship” and define the strategic goals and end points of the endovascular component.

♦ Surgical Management of Arteriovenous Malformations

The most common clinical presentation of iAVMs is hemorrhage. However, in most cases, it is advised to wait until the hematoma begins to resolve, often several weeks after hemorrhage, before definitive surgical resection of an iAVM is initiated. This is a safe strategy, as the risk of early rebleeding is extremely low. Hematomas can often obscure the iAVM nidus, result in fragmentation of the nidus, or promote edema of the surrounding brain parenchyma, which can also further complicate a precise resection. However, the resultant hematoma can occasionally present with life-threatening mass effect and require emergent evacuation ( Fig. 23.2 ). The craniotomy in this situation should be constructed to facilitate either acute or delayed iAVM resection. Normally the initial surgical goal is to relieve mass effect while leaving the iAVM undisturbed, due to the low risk of early rehemorrhage. Re-evaluation of the iAVM should be delayed after evacuation to allow reduction in postoperative edema, resolution of residual hematoma, decompression of the iAVM nidus, and recanalization of previously thrombosed or spastic arterial supply to the iAVM, enabling a more complete determination of the angioarchitecture. A definitive plan can be implemented within 4 to 6 weeks from the initial hemorrhage.

One of the most important aspects in the surgical treatment of iAVMs is the planning of the procedure, which is often multifaceted, entailing a surgical approach, anatomic definition, localization of feeding arteries and draining veins, extirpation, and anticipation of potential pitfalls and complications ( Fig. 23.3 ). The selection of the appropriate surgical corridor should allow the surgeon adequate exposure of the iAVM while maximizing the working space for visualization and lesional manipulation. Brain relaxation and minimization of brain retraction can be facilitated by the neuroanesthesia team (see Chapter 13) with proper head position and cerebrospinal fluid (CSF) drainage. Several different surgical approaches can be utilized depending on the location of the iAVM.

The location of the iAVM nidus can be divided into superficial, lobar, subcortical-deep, and posterior fossa regions. The subcortical-deep locations include the basal ganglia, thalamus, and intraventricular regions. The posterior fossa includes the vermis, cerebellar hemispheres, tonsils, and brainstem. Each location presents specific challenges when approaching, defining, and resecting the iAVM nidus; however, several surgical principles can be applied to all iAVMs irrespective of their primary location.

Surgical Procedures and Considerations

After securing the head in a three-point clamp device, whether radiolucent or opaque, depending on the potential use of intraoperative angiography, strategic positioning of the head can provide several upfront advantages to the operating surgeon. Positioning the head above the heart provides maximum venous outflow as long as the neck has some extension to ensure unobstructed flow through the jugular veins. Further, positioning the head such that the working area sits in a plane that is comfortable for the surgeon to work in during the procedure is paramount (we prefer and recommend that the surgeon be seated during the microsurgical phases of the operation). Localization of the iAVM nidus can be done with a neuronavigation system so that the skin flap can accommodate an adequate bone flap. In the era of minimally invasive approaches, we caution against using too small a craniotomy. Adequate exposure beyond the margins of an AVM is needed to allow safe control of the AVMs vascular supply and venous drainage throughout all phases of the procedure, especially in the unfortunate situation where deep bleeding occurs, resulting in brain expansion and loss of exposure.

Before the skin incision is made, all required surgical instruments, including retractors, aneurysm clips and miniclips, various nonstick bipolar coagulators, the micro-Doppler, and microdissection tools should be available and their functionality checked. In the case of deep iAVMs, long instruments should also be available. After the skin incision is made and the skin flap reflected and protected from potential ischemia, the location of the iAVM should again be localized with the neuronavigation system to ensure proper sizing of the bone flap, which again should be made to encompass the entire area around the iAVM. An adequately sized craniotomy will ensure proper visualization of all surrounding feeding arteries, draining veins, as well as normal brain parenchyma and vasculature. Making the dural opening can be dangerous and should be undertaken with caution, especially in cases of superficial iAVMs where arachnoidal adhesions often bind the iAVM vasculature to the dura. In addition, there are often small dural feeding vessels that should be coagulated and divided. The dural opening should be circumferential and reflected toward a major sinus for protection or in a manner to maximize visualization of the working area. It should also be made some distance away from the cortical aspect of the iAVM if possible, so that important feeding vessels can be preserved and not transected during the opening or the reflection of the dura. Once the dura is reflected, all of the superficial arterial feeders and draining veins should be identified. If the iAVM is located subcortically or deep, neuronavigation should once again be used to ensure that the most direct subarachnoid and trans-sulcal trajectory to the lesion is planned.

The initial phase of dissection is the arachnoid plane ( Fig. 23.4 ) around the cortical arterial feeders, draining veins and surrounding vasculature. Identification of the vasculature is very important as it will define the initial phases of the dissection. For all superficial vessels, each one should be identified as a feeding artery, draining vein, or normal vasculature. Frequently, arterialized draining veins can appear similar to their arterial counterparts. Sometimes neuronavigation can distinguish between these vessels. If there is a question as to whether a particular vessel is an arterial feeder or an arterialized draining vein, further dissection toward the nidus will usually clarify the anatomy. Indocyanine green (ICG) videoangiography ( Fig. 23.5 ) can also assist in this distinction.32 Small vascular clips can also be used to temporarily occlude feeding arteries until their involvement or path to the iAVM nidus can be clarified. Normally, these vessels are highly tol erant of temporary occlusion with minimal effects on the surrounding brain parenchyma; however, the clips should be reopened intermittently to avoid possible ischemic complications in situations where the relationship of the vessel to the nidus is difficult to discern quickly. If the iAVM is located subcortically, superficial vessels can be followed into a surrounding sulcus or fissure toward the iAVM nidus. Once the superficial vasculature is identified and dissected and the margins of the nidus are clarified, the next phase of the dissection can begin.

Once the supplying and draining vessels of the iAVM have been identified and the superficial arterial feeders ligated, a pial-arachnoid plane around the nidus should be carefully established, working in a stepwise circumferential manner, maintaining the patency of the draining veins and carefully coagulating, ligating, dividing, and/or clipping the arterial feeders along the way only at their entry point into the nidus (with care to preserve en passage vessels). Frequently, except in pediatric patients, a plane or rim of gliotic brain tissue will define the boundaries of the iAVM nidus. In this parenchymal phase of dissection, it is very important to dissect the nidus in a spiral fashion, paying careful attention so as not to develop a deeper plane at any point, as bleeding, if encountered, is much more difficult to control and visualize with limited exposure. In other words, “Don’t dig a hole!” ( Fig. 23.6 ).

Once an adequate parenchymal plane is developed, retraction against the iAVM, and not the surrounding brain, can begin. If the drainage system of the nidus has been maintained, the nidus should be a pliable mass. However, the nidus should not be forcibly retracted, as deep arterial feeders are small and friable and can be easily torn resulting in unanticipated hemorrhage. Further, the use of certain liquid embolics can make retraction of the nidus more difficult and less compressible. Polyvinyl alcohol is very compliant, isobutyl 2-cyanoacrylate is rock-like, and n-butyl 2-cyanoacrylate and Onyx (eV3, Inc, Plymouth, MN) are intermediate. Dissection planes can be maintained using either Kendall Telfa (Covidien, Mansfield, MA) or cotton pledgets as the parenchymal dissection proceeds. The dissection plane is continued along the iAVM nidus toward the apex of the lesion. As the nidus is manipulated in an effort to define dissection planes, although always tempting, bipolar coagulation on the nidus itself should be used cautiously as the nidal complex of vessels contains both feeding arteries and draining veins, so that premature coagulation of the main draining may lead to rupture of the nidus. Disruption or coagulation of a significant number of venous loops can increase nidal congestion and lead to premature rupture. Further, translocating feeding arteries from the apex may be torn open leading to difficult-to-control hemorrhaging. If significant hemorrhage does occur during this phase of the resection, the most common cause is that the nidus has been entered ( Fig. 23.7 ). The surgeon should place cotton in the area of hemorrhage and back up to redefine the dissection plane. Compression or retraction of the nidus can sometimes decrease the bleeding enough to localize the source. However, small feeding arteries toward the apex of the nidus are difficult to control due to their retraction into the parenchyma and high pressure of blood supply. Bipolar electrocautery is often ineffective in these locations due to the inability to heat blood in these high-flow situations. Small AVM clips should be used, as they are the most effective tools for hemostasis, as this stops flow, facilitating coagulation and division of the vessel.

If the apex of the nidus abuts an ependymal surface, it is often helpful to open up the ventricle around the lesion so that the deep arterioles can be coagulated against the ventricular wall. The ventricle should be protected with cotton pledgets so that any blood will not track into the ventricular system. As with the superficial dissection, the draining veins, if located deeply, should be maintained until complete disconnection of the arterial supply. As the disconnection nears completion, the turgor in the drainage system should begin to lessen and the appearance of the color should reflect mixed arterialization. Small AVM clips can be systematically applied to the apex of the nidus, and the dissection plane is carried around the deepest aspects of the iAVM. Hemorrhage encountered at this point most commonly represents retained iAVM nidus. If the bleeding cannot be controlled, it is imperative that the iAVM be removed so that the resection cavity can be explored and the source of hemorrhage identified. Before complete nidal extirpation, the turgor of the nidus should be examined after temporary clipping of the final draining vein. If, after several minutes of temporary occlusion, the nidal turgor does not increase, the draining vein should be ligated and divided followed by removal of the iAVM nidus ( Fig. 23.8 ). ICG angiography is also useful to confirm deafferentation.

The resection cavity should be meticulously inspected for complete hemostasis. A Valsalva maneuver can be performed to test for hemorrhage of previously ligated feeding vessels. Although several limitations exist with respect to visualizing residual iAVM nidus within brain tissue, ICG videoangiography can be used to inspect the resection cavity for early venous drainage or retained tufts of arterial feeders. Some institutions utilize intraoperative angiography at this point, which is still the gold standard for imaging residual iAVM nidus. If our index of suspicion is low, we typically maintain careful blood pressure control and perform definitive angiography the following morning.

Related posts:

Carotid Endarterectomy: Decision Analysis and Surgical Technique Moyamoya: Surgical Indications and Strategies Microsurgery for Unruptured Intracranial Aneurysms Cavernoma Revascularization of the Brain Using the ELANA Technique: History, Techniques, Indications, and Future Directions Intracranial Dural Arteriovenous Fistulas

Stay updated, free articles. Join our Telegram channel

Join