Chapter 20: Endovascular Surgical Neuroradiology

Spontaneous rupture of a cerebral aneurysm causing SAH is a life-threatening condition and carries an immediate mortality rate of 10% to 20%.¹ Survival through the initial incident provides no guarantee, however; 12% to 30% of patients fail to recover neurologically, and by 6 months mortality may increase to 50%. Furthermore, as many as one-third of survivors remain dependent following aneurysmal hemorrhage.²

Following initial stabilization and diagnosis, attention necessarily turns toward the prevention of recurrent hemorrhage and vasospasm, both causes of significant morbidity and mortality after the initial hemorrhage.¹ Endovascular therapy is able to address both concerns, and now has an established role in treatment of ruptured cerebral aneurysms. Generally, this is accomplished by endovascular occlusion of the aneurysm with microcoils.

Since the 1960s, it has been known that surgically clipping ruptured aneurysms to prevent recurrent hemorrhage produced results superior to those associated with nonsurgical management. Endovascular coil occlusion of a cerebral aneurysm has been shown to accomplish the same goal: prevention of recurrent hemorrhage. Since its approval in 1991, endovascular coil occlusion of ruptured aneurysms has steadfastly garnered acceptance and has become the treatment modality of choice in patients with SAH due to ruptured aneurysms. Advantages of endovascular treatment include femoral access without a craniotomy and access to midline aneurysms without brain retraction.

The effectiveness of endovascular intervention was directly compared to that of surgical clipping in the International Subarachnoid Aneurysm Trial (ISAT). This trial randomized 2143 patients presenting with SAH attributable to aneurysm rupture and deemed suitable for endovascular treatment or surgical clipping. Patients randomized to endovascular intervention experienced significantly better outcomes. Death or dependence at 1 year in the endovascular group was 23.7% compared with 30.6% in the surgical group. The absolute risk reduction in dependence or death associated with endovascular treatment was 6.9%. After 5 years of follow-up, the dependence rates did not vary significantly between groups; yet, survival in the endovascular group, even after 7 years of follow-up, remained significantly higher. In addition to decreased mortality, endovascular patients also experienced substantially fewer seizures despite a slightly greater incidence of recurrent SAH. After 3258 patient-years of follow-up, there were seven repeat hemorrhages in the endovascular group compared with two rebleeds in 3107 patient-years in the surgical group. This may be related in part to the durability of aneurysm occlusion: Sixty-six percent of follow-up angiograms in the endovascular group demonstrated complete aneurysm occlusion compared with 82% of surgically clipped aneurysms. Meanwhile, 8% of angiograms from endovascularly treated patients demonstrated incomplete occlusion with aneurysm refilling compared with 6% of angiograms of clipped aneurysms.^3,4 Despite lower rates of complete vascular occlusion, endovascular treatment represents a clinically effective treatment for many cases of ruptured aneurysms and is well supported by clinical outcomes as an acceptable treatment method in appropriately selected patients with SAH.

The effectiveness and safety of endovascular intervention was initially demonstrated in 1990 when patients deemed unacceptable surgical candidates were treated with GDC (Guglielmi Detachable Coil, Target Therapeutics, Inc, Fremont, CA) occlusion of their aneurysms.⁵ At that time, reasons to forgo surgical treatment included poor neurologic or medical status, anticipated surgical difficulty, failed surgery, or patient refusal to undergo surgery. Complete aneurysm occlusion was accomplished in 70.8% of small aneurysms with a small neck, 35% of large aneurysms, and 50% of giant aneurysms.^6,7 As a result of procedural complications, 1.74% of patients died, and 4.47% of them died because of the severity of the initial hemorrhage. The GDC procedure received Food and Drug Administration (FDA) approval in 1991, and the use of endovascular coil treatment has increased since that time.

Patients with severe medical comorbidities that could affect their likelihood of withstanding prolonged intracranial surgery are prime candidates for endovascular treatment. Patients with a coagulopathy or those who require chronic anticoagulation are also frequently considered for endovascular over conventional surgery. Evidence derived from multiple retrospective studies has established the safety of endovascular intervention in patients over the age of 70 years with ruptured aneurysms; in such patients, endovascular treatment is generally preferred to surgical clipping.⁸

Clinical grade has a bearing in determining the appropriate intervention. High-grade SAH patients (Hunt and Hess or World Federation of Neurologic Surgeons [WFNS] grades IV and V) present greater surgical challenges, as an edematous or ischemic brain in the setting of increased intracranial pressure (ICP) often responds poorly to surgical manipulation. However, endovascular therapies are less hindered by conditions such as swelling.⁹ A study of endovascular treatment in patients with high-grade SAH due to aneurysm rupture resulted in a favorable outcome in 62% of grade IV and 25% of grade V patients, or in 52.5% of total included patients.¹⁰ A study combining endovascular treatment with aggressive medical management, including hypervolemic hemodilution and hypertensive treatment, in WFNS grade V patients produced encouraging results. A majority of patients (55%) experienced a favorable outcome, while the mortality rate was 18%.¹¹ Thus, there exists growing support for endovascular intervention in patients with high-grade SAH, particularly when coupled with aggressive medical management.

Several specific complications may affect the decision to proceed with surgical or endovascular intervention. The development of an intracerebral hematoma (ICH) following aneurysmal rupture is associated with increased morbidity and mortality compared with SAH alone. The condition may require surgery owing to the need for hematoma evacuation and decompression. Clipping of the ruptured aneurysm may be undertaken simultaneously. Despite evacuation of the hemorrhage, particularly in WFNS grades IV and V, mortality in ICH patients remains high, ranging from 21% to 85%.^12,13

For this reason, a sequential approach involving endovascular aneurysm occlusion followed quickly by surgical evacuation of the hematoma was proposed. One series using this approach in patients with WFNS grade IV or V SAH and ICH due to aneurysm rupture demonstrated favorable outcomes in 48% of patients, with death occurring in 21%.¹⁴ This strategy may be particularly useful when ICH develops opposite to the side of the ruptured aneurysm. Chung et al reported a series of ruptured anterior communicating artery (AComA) aneurysms complicated by significant ICH, making the ICH difficult to evacuate from the optimal site for aneurysm access. These investigators treated AComA aneurysms using endovascular coil occlusion, then evacuated the ICH through a burr hole craniotomy. They observed no rebleed during follow-up and reported that more than half of the patients experienced moderate to good recovery.¹⁵ Thus, evidence suggests that the occlusion of ruptured aneurysms associated with ICH via endovascular means followed by surgical evacuation is a safe and effective approach to a common situation traditionally associated with very high mortality.

Finally, the location and morphology of the aneurysm determine suitability for endovascular treatment. Aneurysms occurring in the posterior circulation are typically better candidates for endovascular treatment based on anatomic considerations. The relationship of these aneurysms to important perforator arteries and cranial nerves creates a great risk of surgical morbidity. Patients with multiple aneurysms in different vascular distributions and SAH of uncertain origin are also well served by endovascular treatment. However, aneurysms with tortuous proximal vessels, severe vaso-spasm, vessels affected by atherosclerosis, or those that are very distally located may present challenges to endovascular access.¹⁶

It is widely accepted that aneurysms with small necks, or inflow regions, relative to fundus size, are particularly amenable to endovascular therapy, while giant and fusiform aneurysms are not without the use of adjunctive devices. This is largely due to the tendency of endovascularly placed coils to herniate into the parent vessel. However, stent-assisted techniques as well as balloon-assisted techniques, which will be discussed in greater detail, help to reduce the failure rate of endovascular treatment in aneurysms with a low dome-to-neck ratio. The relative dimensions of an aneurysm no longer pose an absolute contraindication to endovascular management.

Middle cerebral artery (MCA) aneurysms represent one of the most common sites of intracranial aneurysms, accounting for roughly 20%, and are often managed surgically rather than endovascularly for several reasons. These aneurysms are easily accessible via craniotomy with little brain retraction. Furthermore, MCA aneurysms pose challenges to endovascular treatment, given that they often have wide necks and involve branch vessels. When 53 patients with 58 bifurcation or trifurcation aneurysms of the MCA were evaluated, 88% had a dome-to-neck ratio under 2%, and in 40% branch vessels were incorporated into the aneurysm sac.¹⁷ However, endovascular therapy, while not generally the preferred option, is not precluded in the case of MCA aneurysms. Indeed, a recent study of 16 patients with wide-necked MCA aneurysms, including 10 acutely ruptured aneurysms, treated with stent-assisted embolization found no recurrence, rebleeding, or neurologic deterioration following an average of 20 months of follow-up.¹⁸

Intervention in the care of patients with aneurysmal SAH is primarily undertaken to prevent rebleeding, an event that increases mortality to 70% to 90%.^19,20 The risk of rebleed is greatest in the first 24 hours, reaching 19%, and by 4 weeks the cumulative risk reaches 40%.^1,19 Notably, the risk of rebleeding is greatest in those patients with high clinical grades.¹ Generally, every effort is made to secure the aneurysm as rapidly as possible as soon as the patient is medically stable.

Surgical series have yielded conflicting timing of aneurysm clipping, with some studies demonstrating higher rates of vasospasm and concordant morbidity in patients receiving early surgery, while others demonstrated a benefit associated with very early surgery (0 to 3 days following rupture). Most studies, however, yielded similar results with regard to the generally poorer outcome associated with surgical intervention when the risk of vasospasm is greatest; generally 6 to 10 days postrupture.²¹

Endovascular treatment itself has not been associated with increased vasospasm, and endovascular treatment of vasospasm can be undertaken at the time of aneurysm treatment if necessary. For these reasons, the timing issues involved in conventional surgery do not apparently hold true for endovascular intervention.²² A retrospective study divided patients into three groups depending on when they underwent endovascular coil embolization following the development of aneurysmal SAH: within 48 hours, 3 to 10 days, or 11 to 30 days. Statistically, there was no difference in the percentage of patients experiencing favorable outcomes within each group. There was additionally no difference in the percentage of patients within each group who experienced an improvement in their clinical grade over the course of follow-up.²³ These results suggest that the amount of time elapsed between the rupture and the endovascular intervention has little bearing on the clinical result; nonetheless, treating as soon as possible is recommended given the risk of rerupture.

Endovascular treatment involves placement of prosthetic materials (generally platinum-based coils) within a patient's vasculature to induce thrombosis of the ruptured aneurysm. Excessive thrombus formation could result in vessel occlusion and stroke. Excessive clotting is avoided by the routine use of intravenous heparin to maintain an activated clotting time (ACT) during treatment only. Although there is not universal agreement, many operators aim for an ACT two and a half to three times the baseline value. Following the procedure, the heparin effect can be reversed with protamine sulfate. Patients with small intraparenchymal hematomas have been heparinized safely; furthermore, prior placement of a ventriculostomy is not an absolute contraindication to heparin administration at the time of aneurysm treatment.¹⁶ Patients with experimental use of intra-arterial stents in the presence of SAH, however, may require chronic antiplatelet therapy, a relative contraindication to their use.¹⁶

The patient was likely experiencing cerebral vasospasm, which follows aneurysmal SAH in up to 70% of patients, with 20% to 30% becoming symptomatic and experiencing a delayed ischemic neurologic deficit (DIND).²⁴ This condition remains one of the most challenging events following SAH, with death or permanent neurologic deficits occurring in as many as 20% of symptomatic patients, and generally presents between days 6 and 8.²⁵ Digital subtraction angiography continues to represent the gold standard for diagnosis; however, transcranial Doppler (TCD), CT, magnetic resonance (MR) angiography, and perfusion studies are often used for diagnosis and to guide treatment.²⁴ The medical management of cerebral vasospasm consists of the "triple H": hemodilution, hypertension, and volume expansion. Hypertensive therapy and adequate hydration are now considered the most important components. However, endovascular modalities provide additional options for management of medically refractory vasospasm.

Current endovascular treatment paradigms use intra-arterial administration of potent vasodilators directly into the cerebral arteries and balloon dilatation of narrowed arteries. Although vaso-spasm is primarily a disease of cerebral arterioles, transluminal balloon angioplasty (TBA) using microballoons mounted on microcatheters can be used to dilate the largest cerebral arteries at the skull base. The procedure, which is usually under heparin anticoagulation and often combined with intra-arterial infusion of vasodilators (IAVIs), has been supported by a number of studies that demonstrated clinical improvement in 31% to 80% of patients.^26-29 A meta-analysis evaluating combined TBA and IAVI demonstrated clinical improvement in 62%.³⁰ The procedure has shown benefit whether performed immediately or more than 24 hours following the onset of symptoms; however, evidence associated earlier intervention with better results to prevent extensive cerebral infarction.^27,31,32 Probably the most compelling evidence for the role of endovascular therapy was reported in a large population study by Johnston: At 70 university medical centers offering endovascular therapy, including balloon angioplasty for vasospasm, there was a 16% overall improvement in patient survival.³³ TBA may potentially reduce the risk of vasospasm when applied prophylactically, and in a recent multicenter clinical trial TBA significantly reduced the incidence of DIND; however, benefits must be weighed against the risk of vessel rupture inherent to the procedure, which occurred in 4 of 85 patients undergoing prophylactic TBA in this multicenter trial.³⁴ For this reason, prophylactic therapy is generally not recommended.

Intracranial aneurysms occur in approximately 0.2% to 9.9% of the general population, as determined by autopsy studies.³⁵ Aneurysm rupture is uncommon, yet represents the most devastating sequela of intracranial aneurysms. Other manifestations and symptoms of intracranial aneurysms potentially include severe headaches, mass effect with focal neurologic deficits and cranial nerve palsies. The onset of a third-nerve palsy in association with a growing posterior communicating artery heralds impending rupture. An increasing number of patients are diagnosed with asymptomatic intracranial aneurysms following imaging studies, such as MR imaging (MRI). Indeed, a recent study on incidental findings accompanying MRI found asymptomatic intracranial aneurysms in 1.8% of patients undergoing MRI evaluation.³⁶

Each intracranial aneurysm carries a risk of spontaneous rupture, which is thought to be decreased following endovascular occlusion, although this specific assertion remains unproven. Each intervention carries its own risks, ranging from iatrogenic rupture and intracranial bleeding to thrombosis and resultant stroke.

The International Study of Unruptured Intracranial Aneurysms prospectively followed patients with unruptured aneurysms and found an overall rupture rate of 3%, with 65% mortality in those experiencing hemorrhage; however, the study also established that the rate of rupture varies with size and location. Over a follow-up period extending 5 years, patients with no history of aneurysm rupture and with anterior circulation aneurysms under 7 mm, 7 to 12 mm, 13 to 24 mm, and greater than 25 mm experienced rupture rates of 0%, 2.6%, 14.5%, and 40%, respectively. Patients with a similar background, but aneurysms found in the posterior circulation or the posterior communicating artery, had different rates of rupture: a rate of 2.5% for those smaller than 7 mm, 14.5% for 7 to 12 mm, 18.4% for 13 to 24 mm, and 50% for those larger than 25 mm.³⁷ In addition, those patients who had suffered a ruptured aneurysm in the past proved more likely to suffer rupture of a previously unruptured aneurysm than their counterparts with no history of rupture.³⁸ Additional factors increasing the risk of rupture include severe headaches, tobacco use, and a family history of aneurysmal SAH. Age alone did not have a significant effect, even though cerebral aneurysms are more common with advancing age.¹⁶

The International Study of Unruptured Intracranial Aneurysms followed groups of patients undergoing surgical or endovascular intervention, and facilitated comparisons between the possible approaches to unruptured aneurysms, even though the numbers were too small to achieve statistical significance. Groups undergoing intervention generally contained proportionally more patients with larger aneurysms; however, the endovascular group included a greater proportion of older patients, larger aneurysms (>25 mm), and aneurysms of the posterior circulation as compared with the surgical cohort. The overall morbidity and mortality rate observed 1 year following endovascular surgery was 9.5%, which was slightly less than the 12.15% observed in the surgical group. The study demonstrated that factors, including large aneurysm size and location within the posterior circulation, associated with a greater risk of rupture in the absence of intervention also increased the risk of morbidity and mortality in the groups receiving surgical or endovascular treatment. While age was associated with a poorer prognosis in the surgical groups, it had a less notable impact on the prognosis of those patients treated endovascularly. The study concluded that patients with aneurysms under 7 mm and no history of aneurysm rupture had the lowest risk of rupture, whereas patients with anterior circulation aneurysms under 25 mm experienced the lowest rate of morbidity and mortality associated with interventions.³⁷ Current practice varies widely. Treatment, often in the form of coil embolization, is offered to patients with aneurysms 7 mm or larger. Decisions must be made on a case-by-case basis, however, and treatment for a smaller aneurysm may be appropriate for patients incurring an increased risk of rupture, such as those who smoke, have a family history of aneurysmal SAH, or who have an aneurysm that is symptomatic or demonstrates growth on serial imaging.¹⁶ Patients with a history of SAH and coexisting aneurysms should be considered for treatment regardless of size. In those patients opting to forgo treatment or those with small aneurysms, periodic follow-up with MRI/magnetic resonance angiography (MRA) or CT/computed tomographic angiography (CTA) is advised.³⁹

Endovascular intervention was originally applied to intracranial aneurysms through the use of detachable coils. There are currently six vendors of endovascular microcoils for aneurysm treatment, and each boasts specific proprietary advantages, although none has been proven to be superior over others. After careful angiographic assessment of the aneurysm, a microcatheter is advanced into the aneurysm lumen under fluoroscopic guidance. Subsequently, coils are introduced into the lumen, with the morphology and size of the initial coil selected to match the aneurysm lumen size and to span the aneurysm neck. Aneurysm occlusion is deemed optimal when there is no contrast opacification of the aneurysm lumen. Incomplete aneurysm occlusion may be accepted if there is concern that the placement of additional coils would lead to aneurysm rupture or cause obstruction of the parent artery. Surveillance evaluation after endovascular aneurysm occlusion is commonly performed to assess the durability of treatment. Catheter angiography or MRI/MRA performed at intervals for approximately 18 months is most commonly performed for follow-up.¹⁶

Coil embolization is appropriate for most aneurysms, particularly those with a lumen-to-neck ratio equal to or greater than 2, but is not ideal for fusiform aneurysms, those with vessels originating from the aneurysm, and those with a very large neck relative to parent artery lumen.¹⁶ In a recent study on aneurysms of the posterior circulation, coil embolization succeeded in completely occluding the aneurysm in 80% of cases, while 20% remained incompletely occluded. Complications occurred in 6% and were limited to thromboembolic events causing transient symptoms. Intraprocedural aneurysm rupture represents a serious complication, which is fortunately infrequent, and was found in a retrospective review of 7 years of endovascular intervention to affect approximately 1% of procedures. There is additionally the risk of aneurysm regrowth or recanalization with the need for additional treatment, which affected 7% of patients in that study.¹⁶

Balloon-assisted coiling, which enhances conformation of the coil mass to the shape of the aneurysm, represents an option for wide-neck aneurysms. A temporary balloon is inflated in the parent artery to occlude the neck of the aneurysm while a microcatheter placed into the lumen of the aneurysm is used for coil deposition. The balloon provides a temporary barrier until the aggregate coil mass develops a stable configuration in the aneurysm sac. The technique is associated with occlusion rates of up to 83%, and a series studying its application in wide-necked aneurysms demonstrated a complete occlusion rate of 26% with subtotal occlusion in 53% in difficult aneurysms with wide necks.⁴⁰ However, the procedure is more technically demanding and carries an increased risk of rupture, which may be twice as great as that associated with unassisted coil embolization.⁴¹

Endovascular treatment of wide-neck aneurysms may not be possible even using balloon-remodeling technique. Coronary stents were initially applied off-label in the late 1990s to augment endovascular occlusion of wide-neck cerebral aneurysms. In 2002, the first dedicated neurovascular stent was approved for endovascular occlusion of cerebral aneurysms. Since 2002, several additional devices are now available to treat cerebral aneurysm either by providing a permanent barrier to coil occlusion of cerebral aneurysms or flow remodeling to channel cerebral blood flow away from the aneurysm to induce aneurysm thrombosis and healing. Vascular thrombosis causing stroke remains a major concern following stent placement.⁴² Antithrombotic therapy, usually including use of aspirin and clopidogrel, is used to control the risk of thromboembolism until a neointimal layer covers the prosthetic devices.

Ethylene vinyl copolymer (Onyx, eV3, Inc, Maple Grove, MN) is delivered as a liquid suspension in organic solvent through a microcatheter into an aneurysm but rapidly precipitates to form a solid upon contact with blood. With a balloon inflated over the neck of the aneurysm to prevent extension of the polymer into the lumen of the parent vessel, Onyx is slowly injected to fill the aneurysm. The procedure has been studied in patients deemed poor candidates for coil embolization, as well as in those with aneurysms that recurred or failed to respond to endovascular coil embolization or surgical clipping. Investigators achieved complete occlusion in 79% of aneurysms and subtotal occlusion in 13%. However, there was a risk of delayed occlusion of the parent vessel, which affected 9 of 119 patients.⁴³

AVMs represent aberrant development of the cerebral vasculature. As early as the perinatal period, AVMs may first be recognized as abnormal rests of irregular vessels, or nidi, replacing normal brain. The diagnostic hallmark of an AVM is rapid flow of arterial blood directly into venous structures, bypassing the usual capillary network on a catheter arteriogram. This produces a low-resistance, high-flow system placing high stress on the involved vessels. Consequently, AVMs may be associated with aneurysms along the cerebral blood vessels supplying or draining the AVM. Aneurysms may also occur within the nidi and may be responsible for hemorrhage in some individuals.

The prevalence of AVMs is not known with certainty. An early autopsy study located 30 AVMs among 5754 consecutively performed autopsies, and estimated prevalence at 0.52%.^44,45 However, estimates of prevalence vary, and a subsequent, retrospective, population-based study supported 10.3 per 100,000 as the prevalence of detected intracranial AVMs.^46,47 The New York Islands Arteriovenous Malformation Study was a population study that prospectively followed over 9 million people in the New York metropolitan area and showed an incidence of 1.34 per 100,000 person-years.⁴⁸ While a relatively rare disease, AVMs can cause significant neurologic morbidity and mortality.

AVMs often present with headaches and seizures, occurring in as many as a third of cases, and focal neurologic deficits. Such presentations are theoretically associated with hypoperfusion of the normal brain parenchyma surrounding the AVM as blood flows preferentially through the lower resistance pathways of the AVM called "steal phenomenon"; however, this phenomenon has not been definitively demonstrated in conjunction with symptomatic AVMs. Headaches alone may be experienced by patients with AVMs; however, it should be noted that workups initiated because of the sole symptom of headache yield a diagnosis of AVM in only approximately 0.2% of cases. Furthermore, headaches occurring in the absence of other neurologic symptoms constitute only approximately 0.3% of AVM presentations.⁴⁹ Up to 15% of AVMs are discovered incidentally while asymptomatic; usually during MRI or CT scanning for other reasons.⁴⁵ The most devastating and also one of the most common presentations, in more than 50% of AVMs, involves intracerebal hemorrhage.^50,51

(Figure 20-2)

A variety of conventional imaging modalities are useful in the assessment of brain AVMs and may also lead to the diagnosis of asymptomatic AVMs. As mentioned previously, cross-sectional imaging studies such as CT and MRI are most often acquired first. Catheter angiography permits more accurate assessment of the underlying vascular anatomy of an AVM, including the anatomy of arterial supply and venous drainage, as well as the vascular nidus itself. Conventional MRI and complete catheter angiography provide important information for treatment planning.^51,52

The role of additional assessment tools like functional MRI (fMRI) during treatment planning remains to be determined. fMRI helps to localize the malformation in relation to eloquent parts of the brain. fMRI may be particularly valuable for treatment planning because AVMs appear to cause reorganization of brain function. Likely a developmental process, AVMs may displace functional areas of cortex such that functional centers do not appear in the standard locations.⁵³

The risk of hemorrhage incurred by patients with diagnosed AVMs has been estimated by several studies. Among 139 patients presenting to Columbia University with no history of hemorrhage, the annual risk of hemorrhage was 2.2%.⁵⁴ An annual risk of 2% to 4% for patients with no history of hemorrhage is widely accepted based on the results of several studies following patients not receiving intervention or treatment.^54-56 Notably, the risk profile differs for patients having experienced a previous AVM hemorrhage, as demonstrated by a prospective study with approximately 380 patient-years of follow-up. Within the first year there is a risk of 15.42%; however, this risk decreases to 5.32% over the next 4 years, eventually reaching a plateau at 1.72% after 5 years.⁵⁵ These findings are supported by other studies of AVM populations.^57,58

Factors associated with an increased risk of hemorrhage include infratentorial location (relative risk of 2.65) and deep venous drainage (relative risk 2.50). Young age at presentation (less than 30 years) has also been associated with increased risk, but not at statistically significant levels. Finally, intranidal aneurysms and high intranidal pressures also increase risk. Pressure measurement in feeding arteries of small AVMs demonstrated significantly higher pressures than those of large aneurysms.⁵⁹ In effect, smaller AVMs (diameter ≤ 3 cm) are associated with hemorrhage at presentation significantly more often than larger AVMs.

The Spetzler and Martin (SM) grading scale is a commonly used descriptor and reflects surgical outcomes based on size, location, and venous drainage pattern. This grading scale (Table 20-1) correlates with surgical morbidity and mortality.⁵²

The surgical morbidity for grade I, II, and III lesions is in the range of 1% to 3%. Thus, surgical intervention is recommended for all grade I and II lesions. A case-by-case approach is recommended for grade III AVMs. In grade IV lesions, surgery was associated with 31% morbidity, and the risk climbs to 50% when grade V lesions were operated.⁶⁰ Accordingly, no generalized recommendations were issued regarding grade IV or V AVMs, and it is necessary to individualize the treatment approach. Currently, AVMs are often managed using some combination of endovascular embolization, radiosurgery, and surgery for resection. Treatment resulted in deterioration in 19% of grade I and II patients, 35% of grade III patients, and 42% of patients with either grade IV or V AVMs.^{51, 59}

The association between treatment risk and SM grading is more tenuous when considering endovascular treatment. Following 545 endovascular procedures, 14% of patients experienced new neurologic deficits, 2% of which proved persistent and disabling. Investigators found that the risk of neurologic deterioration following endovascular intervention did not correlate with SM grading, any of the components of the grading system, infratentorial location, the presence of aneurysms, or mode of presentation. Of the variables assessed, only age, lack of neurologic deficit at presentation, and number of embolizations significantly increased treatment risk.⁶¹

A subsequent study based on 295 embolization procedures and reporting good outcomes in 90.5% of patients evaluated outcomes based on the Glasgow Outcome Scale (GOS) and found that those with SM grades III through V, deep venous drainage, and periprocedural hemorrhage had a significantly increased risk of poor outcome.⁶² Similarly, Starke et al defined the correlation of AVM features with endovascular treatment risk in a comprehensive treatment paradigm.⁶³ The resulting AVM Embolization Prognostic Risk Score assigned 1 point for a treatment plan requiring more than one embolization, 1 point for a small-diameter AVM (less than 3 cm), 1 point for eloquent location, 1 point for deep venous drainage, and 3 points for large size (greater than 6 cm). Among 222 patients undergoing embolization prior to microsurgery or radiotherapy, no patients with a score of 0 experienced postprocedural deficits. There was a new deficit rate of 6% associated with a score of 1, a 15% new deficit rate associated with a score of 2, a 21% rate associated with a score of 3, and a new deficit rate of 50% in patients with a score of 5.⁶³

Endovascular embolization of brain AVMs is classically described as an adjunct to surgical resection. In fact, the two liquid embolic agents used to treat cerebral AVMs are specifically approved by the FDA for preoperative embolization. With that said, endovascular embolization of AVMs can be performed using a variety of substances. Embolic agents, including liquids, such as N-butyl-cyanoacrylate (NBCA; Trufill, Cordis Neurovascular, Inc, Miami Lakes, FL), ethylene vinyl copolymer (Onyx, ev3 Neurovascular, Irvine, CA), and undiluted ethanol; as well as particulates, polyvinyl alcohol (Contour, Boston Scientific, Natick, MA), and gelatin spheres (Embospheres, Merit Medical Systems Inc., South Jordan, Utah); and coils from a variety of manufacturers, may be used to occlude an AVM. Each agent has unique advantages and disadvantages. NBCA has received FDA approval and polymerizes to form a solid shortly after exposure to blood. An admixture consisting of NBCA and oil, with the ratio determining the rate of polymerization, is radiopaque and a permanent embolic agent upon polymerization. Furthermore, it causes an inflammatory response in occluded vessels, enhancing its permanence. Particles, often composed of polyvinyl alcohol, are radiolucent and must often be used in conjunction with a radiopaque contrast to provide an indication of placement. Furthermore, recanalization is more common with these particulates than with liquid agents, and the migration of embolic materials into the venous drainage is a theoretical possibility that must be avoided.^51,52

Endovascular embolization can decrease the flow through the AVM nidus and thus reduce the risk of potentially catastrophic intraoperative hemorrhage. Its advantages include reduced surgical time and operative blood loss, and since embolized vessels are easily visualized, preoperative embolization may help with intraoperative navigation. Precise placement of the embolic agents is crucial to prevent stroke or hemorrhage. Embolization of normal arterial branches with a liquid embolic agent risks symptomatic neurologic deficits, which embolization of the venous drainage of an active AVM can create a surgical emergency with hyperpressurization of the nidus and hemorrhage. While feeding artery embolization may be performed with a relatively low risk of complications, embolization of proximal feeder vessels tends to encourage the growth of en passage collateral channels to the nidus, which may complicate surgery with additional bleeding from these friable white matter collaterals.

Preoperative embolization is useful in the treatment of deep-seated lesions. It is used to occlude feeding arteries with associated aneurysms; it can be used as an adjunct to radiosurgery to occlude high-risk features following hemorrhage; and it is often performed in the management of large AVMs (SM grades III to V) with the aim of decreasing surgical risk.⁶⁴ The strategy has proven effectiveness in reducing the size of the nidus. In a recent study using Onyx to preoperatively treat patients with SM grade I to IV AVMs, investigators achieved a mean nidus reduction of 84%. They achieved complete AVM resection with surgery in 98%, with no recurrence after follow-up. Following this combined treatment approach, 38% had a nondisabling neurologic deficit, whereas 7% had a disabling deficit.⁶⁵ In 202 patients who underwent endovascular embolization of AVMs prior to definitive treatment using either surgical resection or radiosurgery at a major academic medical center, Starke et al described a 2.5% risk of morbidity and no mortality.⁶³

Radiotherapy for the treatment of AVMs applies highly focused radiation to the AVM nidus, causing progressive inflammation and vascular sclerosis with resultant thrombosis of the nidus. The advantages of radiotherapy for AVM include its low rate of complications and suitability for patients who are poor candidates for surgical therapy. It is relatively safe for AVMs proximal to eloquent cortex and is most effective for small (3 cm or smaller) AVMs. However, its effects are not immediate and patients remain with a risk of hemorrhage until obliteration of the lesion, which occurs in approximately 80% and is a process that generally takes 2 to 3 years.⁵¹ There are three potential goals of embolization prior to radiosurgery: (1) decreasing the diameter of the nidus to less than 3 cm; (2) eradicating risk factors for hemorrhage such as intranidal and venous aneurysms; and (3) reducing arterial inflow in patients with symptoms, likely due to venous hypertension.⁵¹ Many investigators recommend the use of a relatively permanent embolic agent, such as NBCA or ethylene vinyl copolymer, since less permanent agents such as particulate agents have been associated with higher recanalization rates.^51,52,66 A study utilizing NBCA to preoperatively embolize AVMs in 125 patients, including grades II through VI, determined that embolization produced total occlusion in 11.2% of AVMs and reduced the volume of 76% of AVMs enough to permit radiosurgery, which then produced total occlusion in 65% of partially embolized AVMs. During the latent phase of radiosurgery, partially embolized patients experienced a hemorrhage rate of 3% per year, arguably comparable to the natural history of unruptured AVMs. After 1 year of follow-up, investigators found an 11.8% recanalization rate.⁶⁷ A recent study compared outcomes in patients treated with embolization followed by radiosurgery to those treated with radiosurgery alone. The groups were matched such that postembolization volume was comparable to the volume of unembolized AVMs in the radiosurgery-only group. AVM location and marginal dose of radiation were also equal between the groups. This study showed that patients undergoing radiation alone had a significantly higher likelihood of AVM obliteration (70% compared with 47%). Such patients also experienced better clinical outcomes (64% versus 47%); yet, these differences were not statistically significant.⁶⁸ Although preradiosurgical embolization for the reduction of nidus size did not render the embolized AVM comparable to a naturally smaller AMV, embolization followed by radiosurgery should not be interpreted as an ineffective treatment paradigm. There remains a role for combined therapy to treat larger aneurysms or those that are deep seated and adjacent to eloquent cortex.

Curative embolization of AVMs, defined as the cessation of blood flow through the AVM on catheter angiography, may be achieved through the use of embolization alone. Historically, nidus size has limited the effectiveness of embolization to cure AVMs. In an early study, 71% of AVMs less than 4 mL were fully occluded following endovascular treatment with NBCA, whereas similar results were accomplished in only 15% of AVMs between 4 and 8 mL.⁶⁹ A subsequent study suggested that successful complete occlusion hinged more on the identification of favorable angiographic features, such as a fistulous nidus and a limited number of dominant feeding arteries, than on size. This series reported cure in 74% of favorable AVMs, with an overall cure rate of 40%.⁷⁰ Thus, endovascular cure rates are limited and highly dependent on anatomic features of the AVM.

In selected circumstances, generally where surgical intervention is inappropriate and radiosurgery is not feasible and symptoms, such as seizures, are present, palliative embolization may be cautiously undertaken. In an early study involving 10 patients with AVMs deemed unresectable, 7 of 10 patients improved in terms of seizure reduction and mental status following partial nidal obliteration.⁷¹ Partial obliteration has also improved limb function in lesions proximal to the rolandic cortex.⁷² Any benefits of palliative embolization generally result from the reduction of the vascular steal phenomenon. Steal phenomenon defines the low-resistance blood flow through the AVM that impairs arterial flow to the surrounding normal brain or causes venous hypertension. However, owing to the rapid development of collaterals, clinical improvement stemming from palliative embolization may be transient, and this strategy may ultimately potentially worsen the long-term clinical course by increasing the risk of intracerebral hemorrhage.⁷³

Classically, dAVFs constitute approximately 10% to 15% of all intracranial AVMs. We now recognize that dural fistulas are an etiologically distinct entity. They are different from pial (ie, brain) AVMs in their involvement of the intracranial dural sinuses or cortical veins. There is often no defined core, or nidus, and dAVFs are generally thought to be acquired, rather than congenital, following skull fractures or a thrombotic event in a venous sinus or cortical vein.⁷⁴ The gold standard for visualization of dAVFs remains catheter angiography, although lesions may also be identified through the use of MRI/MRA, contrast-enhanced CT, or CTA. Table 20-2 presents a classification system for dAVFs based upon venous drainage and the amount of shunting, which roughly predict clinical course.⁷⁵

Type I lesions exhibit a benign clinical course. In a clinical series, 0% of type I lesions presented with intracerebral hemorrhage, whereas 2% presented with nonhemorrhagic neurologic deficits. Type II lesions display a more aggressive course, presenting with hemorrhage in 11% and deficit in 39%. The higher likelihood of an aggressive course leads to a recommendation for treatment in most type II dAVFs. Hemorrhagic presentations occurred in 48% of type III lesions, with deficits in 79%.⁷⁵ Again, treatment is recommended for these lesions. Considering all types of dAVFs as a whole, the annual risk of intracranial hemorrhage is approximately 1.5%. This risk increases to 7.4% in patients who have previously experienced intracerebral hemorrhage.⁷⁶ Tentorial, middle cranial fossa, and orbital lesions are associated with more aggressive clinical courses, as are those with cortical or retrograde venous drainage.⁷⁷ Notably, the neurologic symptoms associated with dAVFs depend on the location of the lesion, with cranial nerve syndromes, aphasia, weakness, dysesthesia, dementia, exophthalmos, and visual impairment all within the realm of possibilities.

Current endovascular strategies for the treatment of dAVFs largely resemble the procedures used for AVMs. As is the case for AVMs, only a minority of dAVFs are cured through transarterial embolization. The combination of radiosurgery followed by transarterial embolization, often within 48 hours, yields more promising results. A study involving 105 patients treated in this manner found that at an average of 14 months following the intervention, 68% of patients had fully obliterated dAVFs, whereas 14% had near-total obliteration and demonstrated no cortical venous drainage.⁷⁷ The most significant distinction between the management of dAVFs and AVMs is the option of curative transvenous embolization. Transvenous embolization requires endovascular access to the segment of venous sinus receiving the entire drainage of the fistula. Furthermore, the targeted vasculature must be separate from venous pathways necessary for the drainage of normal brain. If full occlusion of the draining segment is not carefully performed, redirection of venous drainage to delicate cortical veins may further exacerbate the shunt pathophysiology, an event likely to lead to acute hemorrhage.^74,77 Transvenous embolization has demonstrated effectivity in treating the transverse, sigmoid, or cavernous sinuses. In a series of patients treated for dAVFs, all eight patients undergoing transvenous embolization experienced complete occlusion, as compared with five of nine receiving transarterial treatment.⁷⁸ In a large retrospective study of 135 patients with carotid cavernous fistulas spanning 15 years, 76% were treated with transvenous embolization, which came into favor over the course of the study. After an average follow-up of 56 months, 90% of patients were considered clinically cured.⁷⁹ Thus, it is a singularly effective strategy, which must nonetheless be applied cautiously.

Given this patient's presentation, acute stroke is an important consideration. Over 750,000 strokes occur each year; 80% of these are ischemic, while the remainder prove to be hemorrhagic, a distinction that can often be made with the aid of CT scanning.⁸⁰ Specialized imaging modalities, including diffusion/perfusion MRI and multimodal CT, assist in the definition of viable ischemic tissue, although such information does not generally form the basis of treatment decisions (Figure 20-3).

Treatment options pertinent to ischemic stroke include intravenous thrombolytics, intra-arterial thrombolytics, and mechanical embolectomy. Notably, these strategies are contraindicated in the patient with hemorrhagic stroke, in whom they exacerbate hemorrhage and cause clinical deterioration. Currently, intravenous administration of alteplase within 3 hours of stroke onset is the only FDA-approved treatment for acute ischemic stroke. The National Institute of Neurological Disorders and Stroke (NINDS) stroke trial found that significantly more patients (38%) randomized to treatment with intravenous alteplase experienced a complete, or near-complete, recovery after 3 months of follow-up compared with those treated with placebo (21%). While patients treated with alteplase experienced a higher risk of hemorrhage (6.4% versus 0.6%), there was no significant mortality difference between the groups at 3 months.⁸¹ Features such as smaller clot size and more distal location of the clot increase the likelihood of a favorable outcome following fibrinolysis with intravenous alteplase.⁸¹ Owing to this short time window, only 1% to 3% of otherwise eligible patients receive treatment with IV thrombolytics.^82,83

Intra-arterial (IA) fibrinolysis remains unproven, yet it has become the community standard and part of the standard approach to acute ischemic stroke. It has been shown to have a higher recanalization rate, at 63.2%, than intravenous thrombolysis, at 46.2%.⁸⁴ The procedure is performed based on the angiographic visualization of a large cerebral artery occlusion and under fluoroscopic guidance, thus only the amount of thrombolytic agent needed to achieve apparent restoration of flow is used. In the Prolysis in Acute Cerebral Thromboembolism (PROACT II) study, intra-arterial fibrinolysis was tested in a 3- to 6-hour time period from stroke onset. Thus, it is most often used for patients ineligible for IV thrombolysis owing to presentation outside the 3-hour window or for those who fail to respond to intravenous fibrinolysis as determined by the treating physician. Particularly favorable outcomes have followed intra-arterial thrombolysis of occlusions in the internal carotid artery, middle cerebral artery, and basilar artery, whereas thrombi in these locations are generally too large for effective treatment with IV thrombolytics.⁸⁵ A recent meta-analysis demonstrated a significant reduction in the odds of dependency or death associated with IA thrombolysis. Similarly, the randomized, controlled trial known as PROACT II demonstrated increased rates of recanalization and functional independence following intra-arterial administration of recombinant prourokinase compared with that of heparin alone.^86,87 Although patients who received IA prourokinase showed a significantly greater risk of intra-cranial hemorrhage within the first 24 hours, the clinical outcomes in the treatment group were superior to treatment with heparin alone.⁸⁸

Currently, endovascular treatment using various mechanical means to restore cerebral blood flow following acute ischemic stroke is undergoing testing. Embolectomy with or without IA or intravenous fibrinolysis provides an additional option for acute ischemic stroke patients. Recent studies have extended the treatment window to 8 hours. The MERCI Retriever (Concentric Medical, Inc, Mountain View, CA), which has obtained FDA approval, is designed to trap thrombi through inflation of a small balloon distal to thrombus. The device then directs the thrombus into a microcatheter, which is subsequently withdrawn from the patient's body. The MERCI trial demonstrated that mechanical embolectomy significantly increased the portion of patients (48%) experiencing recanalization as compared with historical controls (18%), while demonstrating an association between recanalization and favorable outcome as well as decreased mortality, at least among patients undergoing treatment.⁸⁹ The Multi-MERCI trial introduced a device designed to better trap small thrombi breaking off from the original thrombus, and also included patients not achieving recanalization following attempted IV thrombolysis. Recanalization followed deployment of the device in 55%, while clinically significant procedure-related complications occurred in 5.5% of patients.⁹⁰ Outcome correlated with thrombus location, with the highest rates of recanalization being attained in the posterior circulation and the lowest in the M1 branch of the middle cerebral artery. The procedure appeared safe in patients recently exposed to IV thrombolytics as these patients experienced no statistically significant increase in rates of intracerebral hemorrhage or procedure-related complications. The mortality rate was higher in patients undergoing mechanical embolectomy as compared with control patients in the PROACT II study for IA thrombolysis. However, once only patients satisfying the eligibility criteria of the PROACT II study were considered, the difference ceases to be significant.^90,91

The Penumbra System (Penumbra, Inc, Alameda, CA) represents a newer generation of neurothrombectomy devices designed to remove thrombi causing ischemic stroke from intracranial vessels. It received FDA approval for commercial use based on favorable results in a multicenter study including 125 patients with neurologic defects, who presented within 8 hours of symptom onset and had an angiographic occlusion (TIMI grade 0 or 1).⁹² (TIMI refers to the Thrombolysis in Myocardial Infarction grading system, which was applied to stroke in the PROACT II trial. A score of 0 describes complete occlusion, while a score of 3 refers to complete recanalization.) Postprocedure, 81.6% of patients were revascularized to TIMI scores of 2 (48%) or 3 (52%). Procedural events occurred in 16 (12.8%) and were considered serious in 3 (2.4%). Adverse events, however, were not uniformly attributable to the use of the device. In 35 (28%) patients, intracranial hemorrhage (ICH) was detected on 24-hour CT, but 14 (11.2%) proved to be symptomatic. Although nine patients with detected ICH had received intravenous lytic therapy and three received IA lytic therapy, the investigators did not deem either form of lytic therapy a safety hazard when used with the Penumbra System. Good clinical outcome, defined as improvement by 4 or more points on the National Institute of Health Stroke Scale (NIHSS) score by discharge or a 30-day modified Rankin score (mRS) equal to or less than 2, was achieved by 41.6% of patients. At 90 days of follow-up, 25% of patients achieved an mRS equal to or less than 2; all-cause mortality at this time was 32.8%.⁹² The effectiveness of the device was further supported by a retrospective study of 139 patients meeting device-approval criteria following its release. While 96% of target vessels had TIMI scores of 2 or 3, 84% had scores of 2 or 3 following thrombectomy with the Penumbra System. ICH at 24 hours occurred in 18 patients (13%), and was symptomatic in 10 patients (7.2%). Fifty six percent of patients improved by at least 4 points on the NIHSS. Such results support mechanical embolectomy as a promising treatment modality for ischemic stroke patients.

(Figure 20-4)

Intracranial atherosclerotic disease (ICAD) carries a significant risk and has been recognized as the cause of 8% to 10% of ischemic strokes by population-based and hospital patient–based studies.^93-95 When transient ischemic attacks are considered in addition to ischemic stroke, approximately 100,000 people in the United States experience ischemic events due to ICAD each year. ICAD is associated with diabetes mellitus, smoking, hypertension, and hypercholesterolemia, and can lead to hemodynamic compromise by producing local thrombosis, occluding perforator arteries, or otherwise causing perfusion failure.⁹⁶ It is this hemodynamic compromise, which may be divided into three categories, that ultimately increases the risk of ipsilateral stroke (Table 20-3⁹⁷).

Numerous studies have associated symptomatic intracranial atherosclerosis with a high risk of stroke or death. Patients with symptomatic ICAD, followed in the Warfarin Versus Aspirin Symptomatic Intracranial Disease Study for Stroke (WASID), experienced a 9% to 12% annual risk of ischemic stroke, with 77% of strokes occurring within the first year of follow-up despite optimal medical management and either systemic anticoagulation or antiplatelet therapy. Stenosis of at least 70% increased this risk; such lesions carried a 19% annual risk of stroke in the symptomatic territory.⁹⁸ Similarly, the EC/IC (extracranial/intracranial) Bypass Trial demonstrated that patients treated medically, with management of stroke risk factors and 1300 mg per day of aspirin, for symptomatic ICAD experienced an annual mortality and stroke rate of 8% to 10%.⁹⁹ Annual rates of stroke or death attributable to vascular causes are even higher in those failing antithrombotic therapy, and have been shown to approach 56%.¹⁰⁰

Given the risk associated with symptomatic intracranial atherosclerosis, the failure of medical management to sufficiently mitigate that risk, and lack of a demonstrable decreased risk of recurrent ischemic events following extracranial/intracranial bypass procedures, interventions including angioplasty and stenting have been pursued. Initial procedures were largely corollaries of established techniques in interventional cardiology, utilizing percutaneous coronary intervention balloons to dilate intracranial vessels. However, cerebral arteries differ from their coronary counterparts on several grounds, generally exhibiting smaller diameters, a well-developed tunica media, and a relatively scarce tunica adventitia.⁹⁶ Accordingly, they are more prone to vasospasm and also more likely to rupture at lower forces than are coronary arteries. Furthermore, intracranial vessels may be exceedingly tortuous. Endovascular catheter technology has since progressed, making it an increasingly viable and accepted intervention for patients with ICAD. Currently, patients most likely to undergo endovascular treatment for ICAD remain symptomatic despite optimal medical management with stenosis greater than 50%.⁹⁶ However, decisions regarding individual patients must be made on an individual basis.

Balloon angioplasty has been performed with successful results reported in several case studies. A recent large study followed 120 patients with 124 symptomatic intracranial stenoses treated with angioplasty. Pretreatment stenoses averaged 82.2%, and ranged from 50% to 95%; angioplastic intervention reduced the average to 36%, with a range from 0% to 90%. The investigators observed a combined periprocedural stroke and death rate of 5.8%; after an average of 42.3 months of follow-up, the overall rate of stroke or death was 4.4% annually.¹⁰¹ As expected with any procedure, certain atherosclerotic lesions are more amenable to angioplastic manipulation. Mori and colleagues correlated angiographic features of hemodynamically significant intracranial stenoses with clinical success of intracranial percutaneous transluminal cerebral balloon angioplasty (PTCBA).^96,102

Clinical success rates for types A, B, and C (as described in Table 20-4) were 92%, 86%, and 33%, respectively. Type C lesions were most likely to undergo restenosis, with 100% demonstrating restenosis after 1 year; conversely, no type A lesions restenosed. Risk of fatal or nonfatal ischemic stroke or ipsilateral bypass surgery similarly related to type, with cumulative risks of 8%, 26%, and 87% observed in types A, B, and C, respectively. ¹⁰² Thus, the success of angioplasty depends strongly on the characteristics of the lesion.

Despite the general success of balloon angioplasty, certain complications associated with the procedure, such as elastic recoil and intimal dissection, are averted through stent-supported angioplasty. The Stenting of Symptomatic Atherosclerotic Lesions in the Vertebral or Intracranial Arteries (SSYLVIA) trial introduced the Neurolink stent (Guidant Corporation; Menlo Park, CA), a balloon expandable intracranial stent composed of interconnected bare stainless steel rings and specifically designed to be flexible enough to navigate the tortuous intracranial vessels without causing injury. The multicenter, prospective feasibility study involved patients with symptoms attributable to a single lesion with greater than 50% stenosis. The majority of patients, 43 of 61 (70.5%) had intracranial stenosis, while all others had extracranial vertebral artery stenosis. In 95% of cases the stent was successfully placed, such that less than 30% residual stenosis remained. Within the first 30 days, 6.6% of patients had nonfatal strokes; 7.3% of patients suffered strokes between 30 days and 1 year following stent placement. Restenosis greater than 50% occurred in 32.4% of patients with intracranial lesions; however, 61% of all patients demonstrating restenosis remained asymptomatic, with no strokes or transient ischemic attacks (TIAs).¹⁰³ Although the manufacturers of Neurolink received a humanitarian device exemption (HDE) from the FDA to treat patients with significant atherosclerotic disease, the device is not currently available for commercial use.

The Wingspan stent (Boston Scientific SMART; San Leandro, CA) has also received an HDE and is available to treat patients refractory to medical therapy with symptomatic intracranial stenosis greater than 50%.⁹⁶ It combines balloon dilatation with the subsequent placement of a self-expanding microstent. The study underlying the HDE involved 45 patients with symptomatic intracranial stenosis greater than 50% in a vessel 2.5 to 4.5 mm in diameter, who had been proven to be refractory to medical therapy.¹⁰⁴ Angioplasty and stent deployment was successful in 98% of patients. The 30-day postintervention combined ipsilateral stroke and death rate was 4.5%, whereas the 6-month ipsilateral stroke and death rate was 7.1%. The 6-month rate of all strokes was 9.7%, and all-cause mortality at the same time point was 2.3%. Patients averaged a 74.9% degree stenosis at baseline, which decreased to 31.9% after stenting. At the 6-month follow-up, the average degree of stenosis was 28%. Three patients, however, lost lumen diameter and restenosed to greater than 50% by 6 months of follow-up but were not symptomatic.¹⁰⁴

Since intimal hyperplasia and vascular remodeling theoretically lead to restenosis, stents secreting antiproliferative drugs have been studied as a means of preventing restenosis. Recently, 18 patients were treated with either sirolimus- or paclitaxel-eluting stents.¹⁰⁵ All patients had symptomatic intracranial stenosis greater than 50% despite medical treatment. Within the first month, one major stroke and no deaths occurred; there were no additional such events at the 6-month time point. One of seven patients undergoing follow-up angiography at 6 months had developed restenosis greater than 50%.¹⁰⁵ A larger study also investigating drug-eluting stents detected 1 restenosis from 26 treated intracranial lesions (5%).¹⁰⁶

There exists debate concerning whether angioplasty alone or angioplasty followed by stent placement provides better results. A recent multicenter study compared 95 primary angioplasty procedures against 98 intracranial stent placements.¹⁰⁷ There were no significant differences between the groups in terms of age, gender, or medical history, including hypertension, diabetes, and previous stroke or TIA. Preprocedure stenosis averaged 89.2% and 90.1% in patients undergoing primary angioplasty and those undergoing stent placement, respectively. However, residual stenosis greater than 50% was significantly more common in the angioplasty-treated group, at 15% of patients compared with 4%. The angioplasty-treated group had one periprocedural death and seven periprocedural strokes, whereas the stent-treated group had two periprocedural deaths and seven strokes. These differences as well as differences in terms of stroke- and death-free survival at 2 years were not significant. Similarly, where follow-up data were available, there was no significant difference in the occurrence of restentosis at follow-up: 25 of 66 (38.9%) in the angioplasty-treated group versus 23 of 68 (34%) in the stent-treated group. The investigators concluded that stent placement conferred no significant benefit over angioplasty alone except in the reduction of residual stenosis.

A meta-analysis conducted by Siddiq et al again compared angioplasty and stent placement in patients with symptomatic ICAD.¹⁰⁸ In pooling selected publications, the investigators established a significantly lower rate of death or stroke at 1 year in patients treated with stent placement (14.2%) as opposed to those undergoing primary angioplasty (19.7%). However, periprocedural death and stroke did not differ significantly between the groups. Stent placement was also associated with a lower percentage of restenosis at 11.1% compared with angioplasty alone at 14.2%.¹⁰⁸ These findings indicate that angioplasty followed by stenting may provide benefits over angioplasty alone; however, treatment decisions must be made by experienced interventionalists on a case-by-case basis.

The ischemic brain has increased sensitivity to iodinated contrast. Use of nonionic, iso-osmolar contrast should be limited to 6 mg/kg total, even in patients with good renal function.⁹⁶

In order to prevent reperfusion hemorrhage, a complication that occurs in up to 5% of otherwise successful procedures, mean arterial blood pressure should be lowered by 25% to 30%. Antihypertensive agents that reduce both pressure and pulsatility through α- and β-receptor blockade are ideal unless contraindicated for cardiopulmonary causes. Heparin may be discontinued postoperatively; however, clopidogrel at a dose of 75 mg daily for at least 6 weeks and indefinite aspirin therapy is recommended.⁹⁶

Angiographic follow-up is recommended beginning at 3 months, given the occurrence of 100% restenosis by 3 months in type C lesions observed in a retrospective study.^96,109 Further follow-up should be arranged on a case-by-case basis.