Introduction. Overview

Transcription

1 Interventional neuroradiology in intracranial lesions Robert W Hurst MD ( Dr. Hurst of the University of Pennsylvania Medical Center is a consultant for Boston Scientific Co and a training director for Micrus. ) Matthew Lorincz MD PhD, editor. ( Dr. Lorincz of the University of Michigan has no relevant financial relationships to disclose.) Originally released November 18, 1998; last updated June 15, 2012; expires May 15, 2015 Notice: This article has expired and is therefore not available for CME credit. Introduction Overview In this article, the author updates current endovascular management, particularly emphasizing the role of newer techniques and devices able to more effectively treat increasing numbers of intracranial aneurysms. These advances include the introduction of flow-diverting stents, devices that hold promise to address the still dismal morbidity associated with treatment of large and giant aneurysms. In addition, updated information regarding recently approved devices for endovascular treatment of large-vessel ischemic stroke is becoming available. Retrievable stents, recently approved in the United States, promise to provide higher rates of reopening of acutely occluded large intracranial vessels with possible improvement in overall outcome of these potentially devastating lesions. Key points Dural arteriovenous fistulas are uncommon lesions consisting of direct arteriovenous fistula communications within the dura, usually within the walls of large dural sinuses. Dural arteriovenous fistulas may be characterized by a broad clinical spectrum ranging from asymptomatic to lifethreatening intracranial hemorrhage. Clinical behavior of dural arteriovenous fistulas is most closely related to venous drainage, the feature that forms the basis for both the Borden and Cognard classifications of dural arteriovenous fistulas. Dural arteriovenous fistulas of the transverse and cavernous sinus, the 2 most common locations, are reported to have a benign course in the majority of cases. Dural arteriovenous fistulas involving the tentorium or superior sagittal sinus most often demonstrate aggressive behavior, reflecting their typical drainage pattern into cortical veins. Endovascular techniques currently represent an important and evolving management option for an increasing number of cerebrovascular disorders. The introduction of newer devices for management of intracranial aneurysms has expanded the range of aneurysms that may be effectively treated by these techniques. The availability of mechanical thrombectomy devices increases the effectiveness of reopening large vessels in acute ischemic stroke and holds promise of improved outcomes. Historical note and terminology Interventional neuroradiology uses cerebral angiographic techniques to diagnose and treat disorders of the head, neck, and central nervous system. Cerebral angiography has been of recognized value in the management of cerebrovascular disease since its first performance by Moniz in 1927 (Moniz 1927). With development of Seldinger technique for percutaneous access to the vascular system in 1953, angiography became widely utilized for the diagnosis of vascular disease throughout the body (Seldinger 1953). By 1956, Odman's description of the femoral catheter approach to the innominate and subclavian arteries paved the way for routine angiography of the cerebral vessels (Odman 1956). Interventional techniques using catheters to treat vascular disorders were first applied to vessels outside the CNS and included intentional injection of emboli for vascular occlusion. The development of percutaneous angioplasty for dilation of stenotic vessels by Dotter in 1964 resulted in widespread nonsurgical treatment of atherosclerotic disease throughout the peripheral and coronary circulations (Dotter and Judkins 1964). Nevertheless, the application of interventional catheter techniques to vessels of the CNS lagged considerably because of the small size of vessels involved, the lack of appropriate catheters and devices, and the sensitivity of the brain to ischemic insult.

2 More recently, advances in the development of microcatheters and other devices have extended the availability of interventional radiologic techniques to neuroradiology. Combined with an increased understanding of CNS vascular pathophysiology and improved angiographic technique, interventional neuroradiology has developed a major and increasing role in the safe and effective management of a host of vascular and neoplastic disorders. Currently, interventional neuroradiology techniques should be routinely considered in the management of most vascular disorders of the CNS including aneurysms, vascular malformations, and many types of ischemic stroke. Specific indications for interventional neuroradiology treatment in these entities will be addressed, but it must be recognized that these represent only the most common indications and are constantly evolving. Description Interventional neuroradiologic management of intracranial aneurysms and subarachnoid hemorrhage. Intracranial saccular aneurysms usually occur at vessel bifurcations, typically on the convexity of a curve in the parent vessel, and point in the direction that flow would have continued had the curve not been present. Over 90% of saccular aneurysms involve proximal arteries near the circle of Willis at 1 of the 5 following locations: (1) the junction of the anterior cerebral and anterior communicating artery, (2) the internal carotid artery at the origin of the posterior communicating artery, (3) the bifurcation of the middle cerebral artery, (4) the tip of the basilar artery, and (5) the bifurcation of the internal carotid artery (Rhoton 1980). Aneurysms of the posterior circulation, including those of the basilar tip, comprise approximately 8% to 15% of intracranial aneurysms. In 20% to 33% of cases, aneurysms are multiple. Intracranial aneurysms have a relatively high prevalence of 0.5% to 5% in the population, and their rupture is responsible for nearly 80% of nontraumatic subarachnoid hemorrhages (Horikoshi et al 2002; Winn et al 2002). In the United Sates, aneurysmal subarachnoid hemorrhage affects over 26,000 patients per year with estimates of mortality as high at 60% and significant morbidity affecting as many as half the survivors. Following aneurysmal rupture, the risk of recurrent rupture is particularly high in the first 6 hours after the initial rupture, up to 1.4% on the first day and approximately 1% each day for the first 2 weeks. A high mortality is associated with rebleeding. These figures underline the urgent nature of securing a ruptured aneurysm in a safe and timely fashion. Unruptured aneurysms also have a significant risk of rupture with its associated mortality and morbidity. The risk of rupture for asymptomatic unruptured aneurysms is a subject of great interest in patient management. The International Study of Unruptured Intracranial Aneurysms (ISUIA), originally published in 1998, suggested that some unruptured aneurysms have a lower rate of rupture than had been previously believed (Anonymous 1998). The ISUIA had a retrospective component and a prospective component. The retrospective component examined the natural history of unruptured intracranial aneurysms in patients with unruptured aneurysms. Angiographicallyconfirmed unruptured aneurysms had been selected for followup by investigators at the 53 centers participating in the study. The retrospective component was designed to estimate the cumulative rate of rupture in 2 groups: those patients who had no history of subarachnoid hemorrhage (Group 1) and the patients who had subarachnoid hemorrhage from a different aneurysm that had been successfully repaired (Group 2). The authors reported a relationship between the risk of rupture, the size of the aneurysm, and patient history of subarachnoid hemorrhage from another aneurysm. In 2003, the ISUIA published data on 4060 patients: 1692 did not have aneurysmal repair, 1917 had open surgery, and 451 had endovascular procedures. The investigators found that 5-year cumulative rupture rates for patients who did not have a history of subarachnoid hemorrhage with aneurysms located in internal carotid artery, anterior communicating or anterior cerebral artery, or middle cerebral artery were 0%, 2.6%, 14.5%, and 40% for aneurysms less than 7 mm, 7 to 12 mm, 13 to 24 mm, and 25 mm or greater, respectively, compared with rates of 2.5%, 14.5%, 18.4%, and 50%, respectively, for the same size categories involving posterior circulation and posterior communicating artery aneurysms. The authors indicated that these rupture rates were often equaled or exceeded by the risks associated with surgical or endovascular repair of comparable lesions (Wiebers et al 2003). Considerable controversy continues to surround the meaning of these results (Piepgras 2002; Weir et al 2002; Kailasnath and Dickey 2004). The authors of the ISUIA have suggested that the rate of rupture for most unruptured aneurysms was much lower than had been previously supposed. In contrast, others believe that the results are biased because the study population was selected for followup specifically because the treating physicians at the participating

3 centers believed those patients to have a low risk of rupture. They cite other studies to provide support for rupture rates for selected asymptomatic aneurysms in the range of 1% to 2% per year (Rinkel et al 1998; Juvela 2001). The prospective component of the ISUIA evaluated the mortality and morbidity of elective surgical clipping of aneurysms at the participating centers. The authors found that mortality and morbidity in patients with no prior rupture (Group 1) was related to the patient's age, rapidly increasing for patients over 50 years old. Despite the lower rupture rates and higher surgical mortality and morbidity reported in the ISUIA, treatment of an unruptured aneurysm is associated with much lower mortality and morbidity than aneurysmal subarachnoid hemorrhage. Mortality rates for aneurysmal rupture in the ISUIA were 83% for those patients with no prior subarachnoid hemorrhage and 52% for those with a prior subarachnoid hemorrhage. The ISUIA results emphasize the importance of considering elements of the patient, the aneurysm, and the treatment modality in the individualized decision regarding elective treatment of unruptured aneurysms. Guidelines for decision making in the treatment of unruptured aneurysms incorporating the results of large studies have been published (Komotar et al 2008). The surgical complication rates reported in the ISUIA focus attention on questions regarding optimal management techniques, including endovascular treatment, and outcomes of these potentially devastating lesions. Surgical clipping of aneurysms using open craniotomy had been the primary, and in most cases, only available treatment in the past. The development of and growing experience with detachable coils is supported by an increasing number of studies indicating that many, perhaps most, intracranial aneurysms can be effectively and safely treated using coil embolization with neuroendovascular techniques. Coil embolization of an aneurysm involves angiographic placement of a microcatheter into the lumen of the aneurysm. Detachable coils are advanced through the microcatheter and deployed within the aneurysm lumen. The coil initially remains attached to the guidewire, allowing repositioning to achieve optimal coil placement. The coil is then detached from the wire using one of a number of available detachment mechanisms. Additional coils are advanced through the microcatheter and deployed within the aneurysm lumen until complete filling of the aneurysm lumen is achieved. Post procedure, patients are monitored in an intensive care setting. A key point in approaching current treatment of aneurysms is individualizing treatment. Factors to be considered include those related to the patient, the specifics of the aneurysm, and whether clipping, coiling, or no treatment is the most appropriate course. Patient factors include medical comorbidity including severe lung and cardiac disease. Patients who are elderly, particularly those with poor general medical condition or with decreased life expectancy, also have a higher risk of craniotomy, and they often are treated best with interventional techniques. Aneurysms considered for interventional neuroradiologic treatment must first and foremost include those whose anatomy makes it feasible to occlude the aneurysm lumen while preserving the parent artery. Initially, the vast majority of experience with endovascular treatment of intracranial aneurysms had been with aneurysms having a relatively narrow neck compared with the diameter of the dome of the aneurysm. A number of devices and techniques such as additional coil shapes and materials as well as the use of stents to aid selective placement of coils within the aneurysm lumen have been reported. In addition, recent United States FDA approval of flow-modifying devices promises to improve endovascular treatment results for large, wide-necked aneurysms (Nelson et al 2011). This evolution of endovascular technique has extended the anatomic range of aneurysms amenable to interventional treatment (Han et al 2003; Wanke et al 2003; Benitez et al 2004). A host of features of the aneurysm and the patient must still be considered to determine the relative suitability for each type of treatment, including the aneurysm location and size, the direction in which the dome points, calcification or atherosclerotic involvement of the aneurysm and adjacent vessel, and the presence of intraluminal thrombus. Research also continues to progress with respect to improvement in embolization devices. Use of nonmetallic, coated coils and physiology-modifying devices remains under investigation and promises to increase the effectiveness of endovascular treatment (Kurre and Berkefeld 2008; Klisch et al 2009; White and Raymond 2009). Several benchmark studies of surgical treatment in addition to the ISUIA provide an overview of surgical aneurysm management against which interventional neuroradiologic treatment must be compared. The International Cooperative Study on the Timing of Aneurysm Surgery reported on 3521 patients with aneurysmal subarachnoid hemorrhage, 83% of whom had undergone surgery (Kassell et al 1990a; Kassell et al 1990b). At 6 months, 68% had an excellent or good outcome, whereas 22% were reported as a poor outcome or had died. Miyaoka's series of 1622

4 patients, 80% of whom underwent surgery, reported similar results with 60% excellent or good outcome and 27% poor outcome or dead (Miyaoka et al 1993). A number of studies have further identified subsets of patients or aneurysm locations that have particularly high surgical risks (Kassell et al 1990b). These include patients presenting with poor grade following subarachnoid hemorrhage and those with aneurysms involving the vertebrobasilar circulation. These studies play an important role in the risk benefit determinations as to whether an individual patient should receive surgical or interventional neuroradiology treatment. Several studies have examined the results of coil embolization. One series successfully treated 86% of 203 aneurysms, the majority of which (72%) presented acutely with subarachnoid hemorrhage (Cognard et al 1998). The aneurysms were distributed throughout the anterior and posterior circulations with a particularly high incidence of vertebrobasilar aneurysms (17%). The authors identified total or subtotal occlusion at the time of treatment in excess of 95% of the patients. Eleven percent required retreatment as a result of refilling of the base of the aneurysm with blood. Long-term total aneurysm occlusion was present in 81% of treated patients with a mortality of 2% and morbidity of 4%. The authors indicated that the highest rate of thromboembolic complications occurred in anterior cerebral artery and middle cerebral artery locations and recommend endovascular treatment of these aneurysms only if the anatomy is favorable in the form of a narrow, well defined aneurysm neck. Larger experiences have been published, including a series of 1811 aneurysms (Henkes et al 2004). These data confirm the safety and efficacy of coil embolization for aneurysms throughout the intracranial circulation. Sufficient data are also becoming available to suggest the suitability of various aneurysm locations for different treatments. Aneurysms of the posterior circulation, particularly those at the basilar apex, should be strongly considered for coil embolization as a first line therapy (Gruber et al 1999; Lusseveld et al 2002). In contrast, aneurysms of the middle cerebral artery bifurcation may in many cases be better treated with surgical therapy (Raftopoulos et al 2000). The International Subarachnoid Aneurysm Trial (ISAT) is a major European study evaluating surgical aneurysm clipping verses coil embolization in a prospective randomized trial for ruptured aneurysms (Molyneux et al 2002; Nichols et al 2002). ISAT randomized 2143 patients with ruptured aneurysms to treatment with either neurosurgical clipping or coil embolization. The study found that 23.7% of patients allocated to endovascular treatment were dead or dependent at 1 year compared with 30.6% of those patients treated neurosurgically. This represented a relative risk reduction in dependency or death of 22.6% (an absolute risk reduction of 6.9%) for those treated with neuroendovascular techniques. Most patients in the study had small anterior circulation aneurysms and were in good clinical condition. Rebleeding of target aneurysms occurred in 2.4% of those treated with coil embolization and in 1% of those treated neurosurgically. The study emphasized the fact that patients whose ruptured aneurysms are equally suitable for neurosurgical or neuroendovascular treatment do better when treated with neuroendovascular techniques. It also stressed the need for neuroendovascular consultation in all patients with aneurysmal subarachnoid hemorrhage (Derdeyn et al 2003). Data from ISAT and additional studies supporting acceptable longer-term outcomes for coil embolization of ruptured aneurysms have led to the current AHA recommendation that data support endovascular coil occlusion for patients with ruptured cerebral artery aneurysms that are deemed treatable either by endovascular coiling or by surgical clipping (Meyers et al 2009). With the expanded use of coil embolization for intracranial aneurysms, a number of issues have arisen supporting the need for adjunctive neuroendovascular techniques to enhance long-term permanent occlusion, particularly of larger aneurysms. A major drawback of endovascular aneurysm treatment is the need for retreatment of previously embolized aneurysms due to recanalization of the aneurysm lumen. Recurrence rates of coiled aneurysms are affected by a number of variables involving both the aneurysm and the treatment method. For example, the degree of aneurysm occlusion, length of follow up, and use of adjunctive devices such as stents or balloons are likely to impact recurrence rates. Aneurysm size and neck size clearly affect recurrence, with large and giant aneurysms having recurrence rates as high as 35% to 87%, much higher than rates reported for small aneurysms (Kallmes and Cloft 2007). Giant aneurysms (25 mm or larger) for example, have an extremely poor natural history and remain problematic despite current management techniques. Jahromi and colleagues reviewed long-term clinical and radiological outcome in a series of 39 consecutive giant intracranial aneurysms treated with endovascular repair in 38 patients (Jahromi et al 2008). At follow-up, 95% or higher occlusion rates were documented in only 64% of patients, and 100% occlusion rates were documented in just 36%, with parent vessel preservation maintained in 74%. Twenty percent of treatment sessions resulted in permanent morbidity, and death within 30 days occurred after 8% of treatment sessions. Because

5 1.9 ± 1.1 sessions were required to treat each aneurysm, a cumulative per-patient mortality of 16% and morbidity of 32% occurred. At the last known clinical follow-up examination, 10 patients had worsened neurologic function from baseline (26% morbidity), and 11 had died (29% mortality). The authors state that Although sobering to consider, these rates reflect better outcomes than the previously demonstrated natural history of these lesions and a comparable outcome to open surgical series. Recently developed devices are designed to address some of these considerations by more safely enhancing longterm closure, particularly of large and giant aneurysms. A flow-diverting device was approved in 2011 by the FDA based on data suggesting its ability to effectively close some aneurysms that have been particularly problematic using previously available devices. The Pipeline Embolization Device is a braided cylindrical mesh device with 30% to 35% surface coverage that is implanted across the aneurysm neck and re-lines the diseased vessel. The device has been approved for treatment of adults with large or giant wide-necked intracranial aneurysms in the internal carotid artery from the petrous to the superior hypophyseal segments. Several published studies have evaluated Pipeline treatment of giant or large aneurysms (Lylyk et al 2009; Szikora et al 2010; Nelson et al 2011). Although immediate occlusion occurred in less than 15% of cases, complete occlusion at 6 months ranged from 82% to 94% and was reported at 1 year as from 86% to 94%. Major stroke and neurologic death rates ranged from 0% to 6.5%. Although the small number of clinical studies currently available limits definite conclusions, the device may offer significant benefit to those patients having lesions difficult to treat by other modalities. Perhaps just as important to emphasize is the role of interventional neuroradiology in the management of the most common cause of mortality and morbidity in the survivors of aneurysmal subarachnoid hemorrhage intracranial vasospasm. Interventional neuroradiologic treatment of post-subarachnoid hemorrhage vasospasm. Surgical clipping and neuroendovascular treatment of ruptured aneurysms in the acute phase have minimized rebleeding as a cause of significant morbidity. Nevertheless, the risk of mortality and morbidity associated with subarachnoid hemorrhage does not end with successful closure of the aneurysm. Of those patients well enough to be discharged from the hospital, nearly two thirds never regain the quality of life experienced prior to rupture (Mohr et al 1986). Currently, the leading contributor to both morbidity and mortality in patients surviving aneurysmal subarachnoid hemorrhage is delayed ischemic damage caused by intracranial vasospasm. Although incompletely understood, post-subarachnoid hemorrhage vasospasm appears to be initiated by vasoactive substances resulting from the presence of blood in the subarachnoid space. These substances mediate abnormal constriction of cerebral arteries and narrowing of their luminal diameter. The narrowed lumen potentially decreases delivery of blood to the cerebral parenchyma with consequent ischemia and infarction. However, additional mechanisms likely contribute to delayed ischemic deficits, including abnormal activation of the coagulation cascade, particularly involving the microvasculature; inflammation; and other causes of endothelial and vascular smooth muscle dysfunction. A simplified review of the physiology of intracranial arterial constriction and relaxation highlights a few potential mechanisms contributing to post-subarachnoid hemorrhage vasospasm. Cerebral arterial diameter is mediated by a balance of events causing constriction and dilation of cerebral arterial smooth muscle. The state of smooth muscle tone is significantly dependent on the intracellular calcium concentration. Elevations in intracellular calcium enhance calmodulin-mediated activation of myosin light chain kinase, which in turn phosphorylates myosin to enhance smooth muscle constriction. Decreases in intracellular calcium concentration, including reuptake into the sarcoplasmic reticulum or extrusion through cell membrane channels, act to reverse the constrictive stimulus and cause smooth muscle relaxation and arterial dilation. Endothelial cell production of nitric oxide, a short-lived, highly diffusable molecule produced by the enzyme endothelial nitric oxide synthase, plays a major role in cerebral smooth muscle relaxation. Nitric oxide diffuses from its major site of production in the endothelial cell to the smooth muscle cell, activating the enzyme guanylate cyclase, which increases levels of cyclic GMP. Cyclic GMP, a second messenger, activates a number of guanyl kinase enzymes, which in turn act on myoglobin light chain phosphatases to enhance smooth muscle relaxation. In addition, elevated cgmp levels enhance reuptake of intracellular calcium into the sarcoplasma and closure of voltage-gated membrane calcium channels, thereby lowering intracellular calcium concentrations with resultant relaxation. The state of arterial dilation

6 or constriction is significantly dependent on nitric oxide and smooth muscle cgmp levels, with lowering of either providing strong stimuli to constriction. Degradation of cgmp by one of a family of phosphodiesterase enzymes, primarily PDE5, lowers cgmp levels, enhancing constriction (Dietrich and Dacey 2000; Maxwell 2002; Pluta 2005). Constriction factors have also been described and implicated in post-subarachnoid hemorrhage vasospasm. Endothelin 1 (ET-1), a physiological constrictor molecule produced by the endothelial cell, acts on smooth muscle ETa receptors to elevate smooth muscle intracellular calcium concentration and enhance constriction (Provencio and Vora 2005). A number of components of the complex mixture of subarachnoid blood and its degradation products, including oxyhemoglobin and bilirubin, have been found to disrupt the delicately balanced mechanisms that determine arterial smooth muscle tone, shifting the balance toward constriction. A post-subarachnoid hemorrhage decrease in nitric oxide effect has been identified, including scavenging of the molecule, which likely decreases smooth muscle cgmp, causing constriction. In addition, elevations of smooth muscle calcium concentration and ET-1 have been reported following subarachnoid hemorrhage, perhaps providing a constrictive stimulus (Fassbender et al 2000; Clark and Sharp 2006). Various medications have been used or investigated for the prevention or treatment of post-subarachnoid hemorrhage vasospasm in humans. Administered by various routes, these include calcium channel blockers, nitric oxide donors, phosphodiesterase inhibitors, and endothelin 1 blockers. Calcium channel blockers' effect is potentially mediated by a decrease in smooth muscle calcium levels, decreasing smooth muscle tone. Nimodipine, a calcium channel blocker given orally, is the only medication shown to significantly improve outcome from post-subarachnoid hemorrhage vasospasm. Nimodipine is routinely administered to patients with subarachnoid hemorrhage on admission (Grotta 1991; Feigin et al 1998; Rinkel et al 1998). Although its benefit was initially attributed to a decrease in vasospasm, it has not been shown to affect angiographic vasospasm, raising questions about its site and mechanism of action. Further work has demonstrated an endothelial cell-mediated fibrinolytic effect of this and other dihydropyridine-type calcium channel blockers that may reverse abnormal microvascular coagulation initiated by subarachnoid hemorrhage, which in turn may contribute to the delayed ischemic deficits associated with vasospasm (Vergouwen et al 2007). Nimodipine's antifibrinolytic effect, in addition to effects on calcium channels, may contribute to the clinical benefits. Nevertheless, this research highlights the currently incomplete understanding of the pathophysiology of vasospasm as well as pertinent pharmacologic mechanisms of many available medications. Other medications potentially acting to block calcium channels include nicardipine, a dihydropyridine calcium antagonist, which has also been reported useful in intravenous treatment of angiographic spasm. A beneficial effect on angiographically visible vasospasm has been reported with intraarterial administration of nicardipine into the affected cerebral arterial circulation (Badjatia et al 2004; Hoh and Ogilvy 2005). Intraarterial nicardipine was associated with a low incidence of hypotension and less need for repeat administration than has been reported for papaverine. Nitric oxide precursors such as nitroglycerin and nitroprusside have also been suggested to improve impaired cerebral vasodilator response following subarachnoid hemorrhage. However, both intravenous and intraarterial administration of nitric oxide donors have met with limited success, probably because of features of the metabolic conversion and kinetics of these agents as well as their side effect of severe hypotension (Grasso 2004). Intraventricular administration of nitroprusside has been suggested to be both safe and potentially effective for established vasospasm and cerebral ischemia refractory to conventional treatment (Thomas and McGinnis 2002). Nevertheless, other investigators have suggested considerable variability in both the duration and efficacy of sodium nitroprusside (Mocco et al 2006). Phosphodiesterase inhibitors might also be expected to alleviate vasospasm, presumably by blocking degradation of cgmp and increasing its intracellular concentration. Papaverine is a nonspecific phosphodiesterase inhibitor that acts on both arteries and veins to cause vasodilation. A number of studies have suggested the effectiveness and safety of intraarterial papaverine infusion, most often in conjunction with angioplasty, for the treatment of post-subarachnoid hemorrhage (Kaku et al 1992; Kassell et al 1992; Clouston et al 1995; Mathis et al 1997). Results indicate a 65% to 90% success rate in angiographic dilation of arterial distributions involved with vasospasm. Other studies have associated increases in intracranial pressure with papaverine administration and questioned the efficacy of papaverine treatment alone, even if repeatedly performed, to improve overall outcome (Polin et al 1998; Andaluz et al 2002).

7 Milrinone is an inotropic drug that also dilates vessels by phosphodiesterase inhibition. Arakawa and colleagues examined the effects of intraarterial and subsequent intravenous administration of milrinone in 7 patients with symptomatic cerebral vasospasm. Dilation of the vasospastic vessels occurred in all patients, and cerebral blood flow, calculated in 6 patients, was increased in all. Subsequent intravenous infusion prevented recurrent symptomatic vasospasm in 4 of the 7 patients (Arakawa et al 2001; Arakawa et al 2004; Sayama et al 2006). A number of investigators including Barth and colleagues have evaluated the effects of intravenous clazosentan, a selective endothelin receptor subtype A antagonist, on cerebral perfusion and oxygenation following severe aneurysmal subarachnoid hemorrhage. Their findings suggested that clazosentan may improve cerebral perfusion following aneurysmal subarachnoid hemorrhage (Vajkoczy et al 2005; Barth et al 2007). The complexity of pathophysiology associated with post-subarachnoid hemorrhage vasospasm has so far eluded reliable preventive methods or treatment. In addition, despite reports of successful resolution of vasospasm in case series, effects of medication therapy may be transient, and the results may not be mediated by the most commonly postulated mechanisms. Despite gaps in understanding pathophysiology, vasospasm of the cerebral arteries following exposure to subarachnoid blood appears to be both dose- and time-dependent (Mayberg et al 1990). The distribution and severity of vasospasm have been found to correlate closely with the amount of blood present on early CT scan as well as the clot distribution on CT scan (Adams et al 1987). Other variables, including age, preexisting hypertension, Hunt and Hess grade, and use of cocaine have also been found to determine both incidence and outcome from vasospasm (Howington et al 2003; Rabinstein et al 2003). Clinically symptomatic vasospasm usually presents between the fourth and twelfth day following subarachnoid hemorrhage. New-onset headache or exacerbation of previously present headache is typical, followed by impairment of consciousness, often accompanied by the appearance of focal neurologic deficits. Progression is usually rapid, reaching a peak within a few hours to a day. Estimates of the incidence of angiographically visible post-subarachnoid hemorrhage vasospasm range as high as 76%. More important clinically is symptomatic vasospasm giving rise to ischemic deficits, which is reported in about 30% of aneurysmal subarachnoid hemorrhages. Therefore, clinically significant vasospasm resulting in delayed ischemic deficit is not equivalent to angiographic vasospasm. At present, treatment modalities are directed only toward those patients who manifest symptoms of cerebral ischemia. Transcranial Doppler noninvasively evaluates flow velocities within the major intracranial arteries. Luminal narrowing from vasospasm results in a proportional increase in flow velocity through major vessels as long as constant flow is maintained. Because increased flow velocities precede the development of clinical ischemic deficits, daily transcranial Doppler measurements are extremely useful in identifying high-risk patients prior to the development of clinically significant vasospasm (Seiler et al 1986). Daily transcranial Doppler examinations are currently a standard modality in the evaluation and ICU management of patients with subarachnoid hemorrhage (Aaslid 2002). Therapy for symptomatic patients routinely includes pharmacologically induced hypertension, often combined with volume expansion and hemodilution to increase perfusion to ischemic areas. Reversal of neurologic deterioration may be seen in over 60% of patients so treated. Complications have included rebleeding from unclipped aneurysm, pulmonary edema, hyponatremia, coagulopathy, and myocardial infarction, conditions that may limit the application of medical therapy in some patients. Failure of conventional medical treatment to rapidly resolve vasospasm-induced ischemic deficits should prompt consideration of interventional neuroradiologic therapy including intracranial angioplasty and intraarterial infusion of vasodilating medication. Angioplasty of vasospastic intracranial vessels with an inflatable intravascular balloon catheter results in mechanical dilation of the vessel lumen with concomitant increase in blood flow. First reported in 1984 by Zubkov, multiple series have confirmed the clinical usefulness of intracranial angioplasty in the treatment of selected patients with symptomatic vasospasm (Zubkov et al 1984; Newell et al 1989; Bracard et al 1994; Higashida et al 1994). Persistent neurologic deficits despite maximal medical therapy constitute the major indication for interventional neuroradiological treatment of vasospasm. Prior to angioplasty, CT scan is obtained to exclude etiologies other than

8 vasospasm for new-onset deficits including re-hemorrhage, hydrocephalus, and infarction. Vasospasm in the vessel distribution responsible for the clinical deficit is then confirmed angiographically. Intracranial arteries involved by vasospasm usually demonstrate smooth angiographic narrowing of the vessel lumen. Distal widening of a vessel or a branch whose diameter is greater than the parent vessel also indicates a high probability of involvement by vasospasm. Irregular discontinuous segments may be affected by spasm and, on occasion, a more beaded appearance is present. With severe spasm, angiographically visible delay of flow through the involved vascular distributions may occur. In fact, early delay in cerebral circulation time at the time of aneurysm diagnosis without angiographically visible vasospasm has been shown to be predictive of the later development of clinical vasospasm. This finding supports the involvement vessels too small to be visualized angiographically (Udoetuk et al 2007). Angiographically visible spasm may involve only a localized area adjacent to the aneurysm or be distributed diffusely and extensively throughout the intracranial circulation (Kwak et al 1979). Comparison with prior angiograms is useful to determine the extent of vasospastic narrowing in an individual vessel. Previously present intracranial atherosclerotic disease should be identified as well as areas of vessel hypoplasia. Hypoplasia of the A1 segment is particularly common in patients with aneurysms of the anterior communicating artery complex and must be excluded prior to attempting angioplasty of this segment. Intracranial dilation using angioplasty or pharmacological means is performed only in vessel distributions not harboring an unsecured previously ruptured aneurysm. Increasing blood flow through an unprotected aneurysm may result in hemorrhage with potentially fatal consequences. In cases where a vessel involved by spasm supplies an aneurysm that has not been treated, coiling of the aneurysm can often be accomplished during the same sitting as treatment of vasospasm. Microballoons utilized for intracranial angioplasty of vasospasm are typically inflated with low pressure, thereby minimizing the risk of vessel rupture. It is important to properly select the size of the balloon to avoid overdilation of the vessel, which may lead to rupture. The technique of angioplasty for intracranial vasospasm reflects the underlying pathology of the vessel wall. The mechanism of dilation in vasospasm is believed to be stretching of the smooth muscle component of the vessel wall. This mechanism of vessel dilation is in marked contrast to the mechanism of angioplasty of vessels affected by atherosclerosis. In atherosclerotic disease, relatively higher inflation pressure may be required to fracture plaque and dilate the vessel. In vasospasm, only transient low dilation pressures are necessary to accomplish angioplasty, and high pressures or overdilation of the vessel is to be avoided. Gentle, carefully monitored inflations of the balloon are used to dilate the vessel. Use of a calibrated leak balloon may make the procedure safer by decreasing inflation pressures. A microguidewire is often helpful in vessel navigation. Meticulous use of high quality imaging and roadmapping is essential to ensure safe performance. Intracranial angioplasty is performed in involved vessel distributions responsible for clinical deficits. Dilation using angioplasty is limited to the larger intracranial vessels including the supraclinoid internal carotid artery, intracranial vertebral arteries, basilar artery, and proximal segments of anterior, middle, and posterior cerebral arteries. Attempts to angioplasty more distal segments incur a high risk of vessel rupture with currently available balloon catheters. Following angioplasty for vasospasm, no angiographic evidence of dissection or anatomic disruption of the vessel wall should be present. Restoration of near-normal luminal diameter and configuration is seen throughout the treated portion of the vessel. No recurrence of vasospasm has been observed in treated vessel segments following angioplasty (Yamamoto et al 1992). Post angioplasty angiographic evaluation is performed to insure that all segments of the involved vessel distribution amenable to angioplasty have been dilated. Following angioplasty of proximal vessels, improved filling of the untreated distal portions of the distribution is frequently present. Reported results of intracranial angioplasty indicate that significant clinical improvement may be expected in over 60% of patients treated (Eskridge et al 1998). Complications, the most serious of which is vessel rupture, have been reported in up to 4% of treated patients. The risk of rupture is particularly acute if more distal vessel segments are treated (Newell et al 1989; Higashida et al 1994). Following the procedure, patients are monitored in the intensive care unit and hypertensive hypervolemic therapy continued if necessary.

9 As noted, a number of intraarterially administered medications have shown promise in the treatment of vasospasm. Selective intraarterial administration may be particularly useful when vasospasm involves arterial branches that cannot be selectively catheterized or more distal branches where angioplasty is unsafe. Neuroendovascular treatment offers potential benefit to large numbers of patients with intracranial vasospasm who fail to respond to more conventional medical treatment. These patients would otherwise be expected to sustain severe neurologic deficits, which may result in permanent disability or death. Early consideration and application of these methods is important for the optimal management of appropriately selected patients with post-subarachnoid hemorrhage vasospasm. Nevertheless, additional research is needed to increase the reliability of treatment and improve outcome. Neuroendovascular treatment of intracranial vascular malformations. Vascular malformations of the brain have been classified into 4 major pathological types: (1) arteriovenous malformation, (2) cavernous malformation, (3) capillary telangiectasia, and (4) venous angioma (McCormick 1966; Russell and Rubinstein 1989). Additional lesions considered within this group include arteriovenous fistulas and dural arteriovenous malformations. Estimates of the overall incidence of vascular malformations involving the brain range from 0.1% to 4%. Neuroendovascular therapy of vascular malformations uses embolization to obliterate the nidus or site of arteriovenous shunting. Consequently, only those lesions characterized by arteriovenous shunting are amenable to therapy using neuroendovascular techniques. These include arteriovenous malformations, arteriovenous fistulas, and dural arteriovenous malformations. There is at present no role for neuroendovascular techniques in the treatment of cavernous malformations, venous angiomas (venous vascular malformations), or capillary telangiectasias. The most common clinically symptomatic cerebrovascular malformation, arteriovenous malformation, is estimated to have an incidence of about one seventh that of intracranial aneurysms (Perret and Nashioka 1966). This corresponds to approximately 0.14% of the population. Arteriovenous malformations are congenital anomalies of vascular development resulting in direct communication between arterial and venous channels without an intervening capillary network (Kaplan et al 1961). The focus of therapy in the management of the patient harboring an arteriovenous malformation is the tangle of histologically abnormal vessels representing the site of arteriovenous shunting (the nidus) that replaces the normal arterioles and capillaries with a low-resistance, high-flow vascular bed. The nidus permits increased flow through the arterial feeding vessels to the arteriovenous malformation and delivers increased blood volume under relatively high pressure into the cerebral venous system. Although congenital, arteriovenous malformations most commonly present in the second through fourth decades of life. Intracranial hemorrhage is the presenting symptom in 30% to 55% of arteriovenous malformation patients (Wilkins 1985; Brown et al 1988; Ondra et al 1990). Over 70% of patients who present with hemorrhage do so prior to age 40, most often occurring during the second or third decade. Intracranial hemorrhage associated with arteriovenous malformation is most often intraparenchymal, with the presumed site of bleeding from the nidus or arterialized venous drainage. Intraventricular and subarachnoid hemorrhage may occur as well, although arteriovenous malformations represent the etiology of only a small minority of nontraumatic subarachnoid hemorrhages. The risk of hemorrhage from an arteriovenous malformation has been suggested to be in the range from 2% to 4% per year with a mortality rate of 10% to 15% per episode. In addition, permanent neurologic deficit has been estimated to be 20% to 30% per episode of hemorrhage. Data from uncontrolled studies suggest that arteriovenous malformations rebleed at a rate of approximately 6% in the first year following an episode of hemorrhage. After the first year, the rebleeding rate decreases to that of unruptured arteriovenous malformations, 2% to 4% per year (Perret and Nashioka 1966; Brown et al 1988; Szabo et al 1989). Although considerable, these rates are much lower than the rates of rehemorrhage, mortality, and morbidity associated with aneurysmal hemorrhage, possibly as a result of the subpial location of arteriovenous malformations and lower pressures associated with arteriovenous malformation hemorrhages. Importantly, there may be a significant spectrum of hemorrhage rates based on specific clinical and anatomic features of individual arteriovenous malformations. Stapf and colleagues in an analysis of data on 622 consecutive patients with arteriovenous malformation found that annual hemorrhage rates on follow-up ranged from 0.9% to as high as 34.4% (Stapf et al 2006). They determined that hemorrhage rates were lower in patients without hemorrhagic arteriovenous malformation presentation, deep arteriovenous malformation location, or deep venous drainage and were highest in patients with all 3 of these risk factors. Accurate data on natural history of brain arteriovenous malformations are

10 obviously essential for informed decision making. To more completely address this question, A Randomized Trial of Unruptured Brain AVMs" (ARUBA) was begun in 2007 and is currently underway (Mohr 2008). ARUBA is a randomized controlled clinical trial of adults with the diagnosis of an unruptured brain arteriovenous malformation. Patients willing to participate are randomly assigned to either best possible invasive therapy or medical management without intervention. The trial is designed to evaluate the benefits of intervention versus conservative (medical) management for arteriovenous malformations that have not yet bled (Stapf 2010). Seizures are also a common clinical manifestation of arteriovenous malformations, reported as a presenting symptom in 20% to 60% of several large series (Perret and Nashioka 1966; Waltimo 1973; Graf et al 1983; Mendalow et al 1987). Acute or progressive neurologic deficits may result from intracranial arteriovenous malformations even in the absence of acute hemorrhage. Presumed mechanisms include "steal" of blood from adjacent normal brain into the lowpressure, high-flow vessels feeding the arteriovenous malformation. Dilation of the enlarged draining veins may result in mass effect with resultant compression and neurologic dysfunction. Venous hypertension represents an additional mechanism of neurologic dysfunction that may affect brain adjacent to or at a distance from the arteriovenous malformation nidus. Hydrocephalus may develop either as a result of prior hemorrhage or by compression of adjacent CSF pathways and represents an additional possible mechanism of neurologic dysfunction. Clinical applications Because of the relatively low rebleeding rate associated with hemorrhage from arteriovenous malformation, acute or emergent intervention is usually limited to patients with life-threatening mass effect from intracranial hemorrhage. Timing of treatment is determined by the clinical presentation and by the anatomy of the arteriovenous malformation. Traditionally, treatment for arteriovenous malformation had been complete surgical excision of the nidus, which can be achieved in approximately 80% of patients with a mortality and morbidity better than the natural history. Depending on the risk associated with surgical resection alone, adjunctive or alternative forms of therapy may be employed. Because arteriovenous malformations may have a grim prognosis if recurrent hemorrhage occurs, the goal of management regardless of the treatment modality is obliteration of the nidus for cure. Neuroendovascular techniques contribute significantly to the management of the arteriovenous malformation patient (Smith and Garg 2002). Embolization of arteriovenous malformations using interventional neuroradiology techniques has been found useful in 4 general management situations related to arteriovenous malformations: (1) Preoperative embolization may decrease nidus size and flow, decreasing intraoperative hemorrhage and, therefore, the risk of surgical resection. In addition, specific feeding pedicles can often be targeted for closure using interventional neuroradiology techniques. These may be pedicles arising from adjacent vascular distributions, those with a particularly large arteriovenous shunt, or those feeding intranidal aneurysms that may increase the risk of hemorrhage. (2) Embolization has also been shown to have a significant role in decreasing the size of the arteriovenous malformation nidus prior to radiosurgical therapy. Radiosurgery, or stereotactic external beam radiation therapy, uses focused beams of radiation to give high doses to an arteriovenous malformation nidus while minimizing exposure to the surrounding brain. The technique causes obliteration of the nidus secondary to radiation damage. Radiosurgery is usually pursued in cases considered unsuitable for resection, due to either location of the arteriovenous malformation or overall operative risk. Obliteration rates in the range of 70% to 81% within 2 years of treatment have been reported with permanent neurologic complications in the range of 3% to 10% (Steiner 1986; Lunsford et al 1990; Ogilvy 1990; Fabrikant et al 1992). Higher obliteration rates have, however, been documented with smaller arteriovenous malformations in a number of studies (Friedman et al 1995; Pollock et al 1996). Data have emphasized the usefulness of this modality in the treatment of deep central arteriovenous malformations, a particularly difficult subset of these lesions (Andrade-Souza et al 2005). Because the maximum size of the target diameter in radiosurgery is in the range of 3 cm, larger arteriovenous malformations may not be initially amenable to radiosurgical treatment. In such cases, arteriovenous malformations can be reduced in size and radiosurgery made possible or more likely successful through the use of embolization with a permanent embolic agent. Techniques of staged radiosurgery suggest that this technique may have a role in the treatment of even larger arteriovenous malformations than has been routine in the past (Pendl et al 2000). (3) In approximately 10% of cases, complete obliteration of arteriovenous malformations can be achieved by embolization using permanent embolic agents, making further therapy unnecessary (Berenstein and Lasjaunias 1992). (4) Lastly, partial embolization of arteriovenous malformations responsible for progressive neurologic deterioration may suggest a role for neuroendovascular techniques in palliative treatment of some patients. In most cases, these patients have extensive arteriovenous malformations not amenable to complete obliteration by even combination therapy. Several studies have documented remission of neurologic deficits following partial embolization.