The classic definition of a transient ischemic attack (TIA) is a sudden focal neurologic deficit, referable to a specific arterial distribution and of presumed vascular origin that resolves within 24 hours. However, the majority resolve within 1 hour, and only 15% of those persisting beyond 1 hour resolve within 24 hours. Neuroimaging evidence of cerebral infarction can be demonstrated in 15% to 20% of computed tomography (CT) scans and 50% to 71% of diffusion-weighted magnetic resonance imaging (DWI) with symptoms of duration greater than 6 hours but less than 24 hours. The perception that TIAs are benign whereas strokes are more serious results in inadequate patient reporting and physician treatment (2). In two recent studies of patients presenting with TIAs, 8% to 10% suffered a stroke. However, there is a high-risk subset of patients with TIAs whose risk of stroke is 10% within the first week, and for these patients, emergent evaluation and treatment is essential, particularly because less than 20% of strokes are forewarned by TIA symptoms (3). High-risk features that were identified include older age, hypertension, focal symptoms consisting of unilateral weakness or speech disturbance, symptom duration, and diabetes. The “ABCD” risk scoring system separated those patients who did not go on to have an early recurrent stroke from a group consisting of 95% of the patients who had a 30% risk of having a stroke within the first week (4). These data can be used to determine which patients may require hospital admission for urgent evaluation and treatment (Table 107.3).

Hypertension is the most important risk factor for both ischemic and hemorrhagic stroke, and given its prevalence, it should be a major focus of therapy. The risk of stroke is directly related to the magnitude of elevation of both the systolic and
diastolic blood pressures for both genders and all age groups. The lifetime stroke risk for those with normal blood pressure (<120/80 mm Hg) is half of that for individuals with hypertension. Isolated systolic hypertension increases stroke risk two to four times even after controlling for age and diastolic blood pressure.

The decline in stroke mortality has been credited to the effective treatment of hypertension. From a meta-analysis of 14 trials of hypertension therapy, a 35% to 40% reduction in stroke risk over 5 years was obtained with a 5- to 6-mm Hg reduction in diastolic blood pressure. The guidelines of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC 7) define normal blood pressure as less than 120/80 mm Hg, but only 1/2 of hypertensive individuals even achieve the treatment goal of less than 140/90 mm Hg. The Perindopril Protection Against Recurrent Stroke Study (PROGRESS) compared perindopril versus placebo in 6,102 patients with TIA or stroke within the prior 5 years (5). Of note, more than half of the patients had blood pressure (BP) less than 130/85 mm Hg at entry. Combination therapy with indapamide was administered with physician discretion, and nearly half of the patients were taking additional antihypertensive therapies. In follow-up averaging 4.2 years, combination therapy significantly reduced the risk of cerebral hemorrhage by 50% and ischemic stroke by 24%, with average BP lowering of 12 mm Hg systolic and 5 mm Hg diastolic, suggesting that a more aggressive target is warranted. Because the stroke risk reduction varies by stroke subtype, it is likely that different BP targets will be appropriate for different stroke subtypes. Along the lines of the more aggressive JNC 7 target for patients with small-vessel disease with diabetes or renal insufficiency, the Stroke Prevention in Small Subcortical Strokes trial (SPS-3) is investigating a systolic BP target of less than 130 mm Hg versus 130 to 149 mm Hg in patients with a prior lacunar stroke due to small-vessel disease.

The confusion regarding hyperlipidemia as a risk factor for stroke lies in the heterogeneity of stroke subtypes.
Hypercholesterolemia is a powerful risk factor for ischemic stroke due to atherothrombosis; however an increase in fatal hemorrhage stroke is related to low cholesterol levels (6). In clinical trials of secondary prevention of coronary heart disease (CHD) in which stroke was a prespecified endpoint, statin therapy reduces stroke by 23% to 31%. In trials of primary prevention and those including patients with cerebrovascular disease without established coronary heart disease, the low event rate for stroke underpowered an analysis. In a meta-analysis of more than 200,000, largely low-risk, patients, statin therapy reduced the risk of stroke by 25% in patients with or without CHD, which translates into a number needed to treat (NNT) of 2,778 patients for 1 year to prevent 1 stroke. Benefit was not identified for other lipid-lowering strategies, including diet (7). In a higher-risk subgroup of patients with a prior TIA or ischemic stroke but without CHD from the Heart Protection Study, statin therapy reduced the risk of recurrent stroke by 21, which translates into an NNT of 102 patients for 1 year to prevent 1 stroke. Current recommendations are to initiate statin therapy in patients with symptomatic carotid atherosclerosis as a high CHD risk-equivalent and after a recent TIA or ischemic stroke. The Stroke Prevention by Aggressive Reduction in Cholesterol Levels (SPARCL) trial comparing atorvastatin 80 mg/day versus placebo initiated 1 to 6 months after a TIA or ischemic stroke should address the role of intensive therapy in this targeted population (8).

The incidence of type 2 diabetes has increased dramatically and is a major risk factor for stroke, and hyperglycemia in the setting of an acute stroke is associated with poorer outcome with increased infarct size. Strict blood sugar control has been shown to reduce microvascular complications, although it is not clear that this extends to lacunar infarction of the brain from small-vessel disease. Although aggressive treatment of hypertension among type 2 diabetics has been shown to reduce stroke risk, tight diabetic control has not. Hyperglycemia in the setting of acute stroke is associated with poorer outcome related to increased infarct size. Insulin resistance and the metabolic syndrome have been demonstrated to have an intermediate risk for stroke when compared to diabetes or controls. The ongoing Insulin Resistance Intervention after Stroke (IRIS) trial will test the efficacy of insulin sensitizer therapy with pioglitazone to prevent recurrent stroke and other vascular events in nondiabetic patients with insulin resistance who have had a recent ischemic stroke.

The risk of stroke increases with the severity of stenosis from 1%/year to 2%/year for less than 75% stenosis to 3.3%/year for greater than 75% stenosis, of which 2.5% represent infarcts within the territory of the stenotic artery (9). However, carotid bruits and carotid plaque thickness are stronger indicators of the risk of myocardial ischemia and vascular death, at 8%/year to 10%/year. In the 1980s, a Rand report suggested that one third of carotid endarterectomies were performed for inappropriate indications, and an additional one third were questionable. This prompted the organization of multiple randomized clinical trials to investigate the role of carotid revascularization procedures versus medical therapy for stroke prevention (Table 107.4).

The Carotid Artery Stenosis with Asymptomatic Narrowing: Operation Versus Aspirin (CASANOVA) study (11), the Veterans Administration Asymptomatic Carotid Stenosis Trial
(12), and the Mayo Asymptomatic Carotid Endarterectomy (MACE) trial (13) were negative or inconclusive, and the MACE trial was halted due to an excess of myocardial infarctions (MIs) in patients not on aspirin therapy. The Asymptomatic Carotid Atherosclerosis Study (ACAS) (14), which randomized 1,662 patients with greater than 60% internal carotid artery stenosis (determined by ultrasound or angiography), was stopped after 2.7 years when the projected 5-year stroke rate was significantly reduced from 11.3% to 5.6% with surgery performed at a low, 3% morbidity and mortality rate. Women did not benefit from the procedure, partly due to their higher rate of angiographic complications.

The North American Carotid Endarterectomy Trial (NASCET) randomized patients with a recent (within 3 months) ipsiterritory TIA or nondisabling ischemic stroke and 30% to 99% stenosis defined by angiographic measurements of the internal carotid artery (15). The “NASCET method” of measurement has become the standard for describing carotid stenosis and predicting a clinical benefit with therapy as described by these outcomes-driven clinical trials.

As the severity of carotid stenosis increased, so did the risk of having a stroke; as stroke risk increased, so did the benefits of surgery. Factors that identified those at high risk of stroke with medical therapy alone include age greater than 70 years, male gender, systolic BP greater than 160 mm Hg, diastolic BP greater than 90 mm Hg, symptoms within 31 days, history of stroke, stenosis greater than 80%, plaque ulceration, and a history of either smoking, hypertension, MI, congestive heart failure, diabetes mellitus, claudication, or hyperlipidemia. Those with six or more risk factors had a 39% risk of ipsilateral stroke within 2 years.

The severe (70% to 99%) stenosis subgroup was prematurely terminated when surgery significantly reduced the risk of any ipsilateral stroke at 2 years from 26% to 9% (65% relative risk reduction) and major or fatal ipsilateral stroke from 13.1% to 2.5% (81% relative risk reduction). For those with 50% to 69% stenosis, surgery also reduced the rate of any ipsilateral stroke from 22.2% to 15.7% (29% relative risk reduction) at 5 years, but the benefit was marginal (p = .045). Within this subgroup, characteristics that were associated with a greater benefit from surgery included male gender, a recent stroke (rather than a TIA), recent hemispheric symptoms (rather than retinal symptoms), and failing aspirin at 650 mg or more daily. For those with 30% to 49% stenosis, there was no significant benefit in reducing ipsilateral stroke at 5 years with surgery (14.9%) compared to medical therapy alone (18.7%). The highest risk for stroke was immediately after the index ischemic event and declined to 3%/year within 2 to 3 years; if the patient escaped recurrent symptoms during that time, he or she had little to gain from having delayed surgery. These results were confirmed in the European Carotid Surgery Trial (ECST) (16).

The benefits of carotid endarterectomy surgery were realized when the perioperative morbidity and mortality was 5.8%, which included a major stroke and death rate of 2% and a mortality rate of less than 1%. Characteristics that doubled the risk of perioperative stroke or death included contralateral carotid occlusion, evidence of an ipsilateral cerebral infarct on CT/MRI, left-sided carotid disease, diabetes, diastolic blood pressure above 90 mm Hg, absence of a history of myocardial infarction or angina, and taking less than 650 mg of aspirin per day, but not age or gender.

Guidelines for endarterectomy advise that the best indication for carotid endarterectomy is for the prevention of ipsilateral carotid territory ischemic stroke in patients with a recent transient ischemic attack or minor ischemic stroke due to greater than 70% ipsilateral carotid stenosis and is also beneficial for those with 50% to 69% stenosis if they have a good 5-year life expectancy and the procedure can be performed with a combined morbidity and mortality of less than 6%. For asymptomatic patients with greater than 60% stenosis, surgery should be recommended if the patient has a good 5-year life expectancy and the procedure can be performed with a combined morbidity and mortality rate of less than 3%. All patients should receive treatment of risk factors, patient education about TIAs, and antiplatelet therapy.

Endovascular approaches to carotid revascularization were developed in part because the low perioperative risks demonstrated in the randomized trials did not reflect clinical practice, with mortality rates of 1.4% for patients not randomized at trial hospitals to 2.5% for patients treated at low-volume, nontrial hospitals (17).

The Carotid and Vertebral Artery Transluminal Angioplasty Study (CAVATAS) was the first large-scale trial comparing carotid angioplasty, 26% with stents, to carotid endarterectomy (18). At 30 days, the 10% risk of death or stroke did not differ between the endarterectomy and angioplasty groups, but the 5.9% risk of death or disabling stroke with endarterectomy was substantially higher than the rate of less than 3% reported by NASCET and ECST. Registry data tracked the improvements in endovascular revascularization with carotid stenting, particularly with emboli protection devices. In a report of 11,243 patients entered into a global stenting registry, the 30-day rate of stroke rate was 5.3% without and 2.2% with emboli protection (19).

The Stenting and Angioplasty with Protection in Patients at High Risk for Endarterectomy (SAPPHIRE) trial was a multicenter, randomized trial of carotid stenting versus endarterectomy in asymptomatic greater than 80% or symptomatic greater than 50% carotid stenosis in patients deemed to be of high surgical risk but eligible for either procedure, with a companion registry when patients were not (20). All patients had at least one high-risk feature consisting of age greater than 80 years, significant cardiac or pulmonary disease, contralateral carotid occlusion or laryngeal-nerve palsy, prior neck surgery or radiation, or restenosis after carotid endarterectomy. All stenting patients had emboli protection (Angioguard or Angioguard XP; Cordis, Miami Lakes, FL) before deployment of a nitinol stent (Precise or Smart stent; Cordis). The study enrolled 747 patients, 334 randomized, 406 in the stent registry, and only 7 in the surgical registry. The primary endpoint of death, MI or stroke at 30 days plus death from neurologic causes or ipsilateral stroke from 31 days to 1 year, was significantly lower for carotid stenting compared to endarterectomy (12.2% vs. 20.1%, p = .05), with trends to decreased myocardial infarction (3.0% vs. 7.5%, p = .07) and death (7.4% vs. 13.5%, p = .08). Target vessel revascularization was significantly lower for carotid stent patients (0.6% vs. 4.3%, p = .04), as was cranial nerve palsy (0% vs. 4.9%, p = .004).

Multiple additional device registries have reported similar risks of death, MI, or stroke at 1 year from 4.5% to 9.1% by various specialties and including community hospitals. On March 17, 2005, the Centers for Medicare and Medicaid Services (CMS) expanded the reimbursement of carotid stenting to high-risk patients with symptomatic greater than 70% carotid artery stenosis. CMS will also reimburse carotid stenting for high-risk patients with greater than 80% stenosis if they are enrolled in clinical trials pending the results of several ongoing randomized trials comparing stenting to endarterectomy that have agreed to combine data for a meta-analysis. In the United States, the National Institutes of Health (NIH)-sponsored Carotid Revascularization Endarterectomy Versus Stent (CREST) trial is enrolling symptomatic and asymptomatic patients, and the Asymptomatic Carotid Stenosis Stenting Versus Endarterectomy Trial (ACT-I) is randomizing asymptomatic patients with severe carotid artery stenosis to either stenting or endarterectomy in a 3 : 1 ratio.

Intracranial atherosclerosis accounts for about 10% of all ischemic stroke and is a high-risk condition. The risk of recurrent stroke with symptomatic greater than 50% stenosis is 19%, with 1/4 occurring in territory of the symptomatic artery, often early within the first year, and at similar rates regardless of oral anticoagulant or antiplatelet therapy (21). The risk of stroke with intracranial stenting was 11.5% within the first year in 61 patients treated with Neurolink stent in the Stenting of Symptomatic Atherosclerotic Lesions in the Vertebral or Intracranial Arteries (SSYLVIA) registry (22) and 7.1% within 6 months in 45 patients treated with the Wingspan stent (23). Because the rate of restenosis greater than 50% at 6 months was lower with the Wingspan stent (7.5% vs. 35%), it was chosen for an NIH phase I safety trial in patients with symptomatic 70% to 99% intracranial stenosis. Symptomatic intracranial atherosclerosis is also a predictor of high CHD risk, and the rate of MI or sudden death within 2 years was significantly higher in warfarin-treated versus aspirin-treated patients in the Warfarin-Aspirin Symptomatic Intracranial Disease (WASID) trial, 7.3% versus 2.9% (21).

The presence of cardiac disease doubles the risk for stroke. Concomitant coronary heart disease (CAD) is frequent in patients with cerebrovascular disease; 86% have angiographic evidence, of which 40% are severe, and for patients with atherothrombotic stroke, MI is the most frequent cause of death. Cardiogenic embolism accounts for approximately 15% to 20% of all ischemic strokes. The common clinical syndromes are those involving the middle cerebral artery, posterior cerebral artery, or top of the basilar and reflect cerebral blood flow patterns. Some causes of cardiogenic cerebral embolism are well established; other more commonly encountered cardiac conditions are less well established. More detailed discussions are given elsewhere in this text.

Overall, 1/2 of all cardioembolic strokes and one-third of strokes occurring in the elderly are in the setting of atrial fibrillation (AF). One-third of patients with AF sustain a stroke during their lifetime, and one third have evidence of a silent stroke. The perception that oral anticoagulation was hazardous limited its use until pivotal clinical trials in the 1980s and the development and widespread adoption of the International Normalized Ratio (INR) method of anticoagulant monitoring. Stroke risk reduction with warfarin was 68% (range 50% to 79%), which increased to 83% when target anticoagulation of INR 2 to 3 was achieved. Failure of anticoagulation to prevent stroke was associated with an INR close to 1.0 in 60% of patients. The stroke risk reduction with aspirin alone, particularly 325 mg/day, was 20% to 25% when compared to placebo. Recommendations for antithrombotic therapy balance the expected benefit of protection from thromboembolic stroke and systemic embolism against the risk of intracranial hemorrhage and systemic bleeding.

The nonvalvular atrial fibrillation (NVAF) trials represent the most extensive data set for stroke prevention focused on a very specific stroke subtype, in contrast to most of the stroke prevention literature, which include many stroke subtypes with variable natural histories. As a result of these outcomes studies, calculators have been developed to further subclassify patients by risk and guide recommendations for antithrombotic therapies. The recommended therapy for patients at high risk and most patients at intermediate risk is warfarin with an INR range 2.0 to 3.0 for low-risk aspirin. Combination therapy adding low-dose aspirin to standard warfarin therapy should be considered for those with significant coronary heart disease risk, again balancing potential benefits against the increased risks of bleeding with combination therapy (24).

Despite these data, 50% to 75% of patients with atrial fibrillation do not receive appropriate treatment. The most frequently cited reasons are the fear of hemorrhagic complications, particularly hemorrhagic stroke, and the requirement for regular monitoring and adjustment of therapy. Aggregate data on the complications of oral anticoagulant therapy for NVAF included intracerebral hemorrhage at an overall rate of 0.3%/year, although patients were excluded from the trials if they were at risk for bleeding including prior intracranial hemorrhage, predisposition to trauma, inability for adequate follow-up, uncontrolled hypertension, or alcohol abuse. In warfarin-eligible patients, standard warfarin therapy was superior to the combination therapy of aspirin with low-intensity, fixed-dose warfarin (INR 1.2 to 1.5) in the Stroke Prevention in Atrial Fibrillation III (SPAF III) trial and combination aspirin with clopidogrel 75 mg/day in the Atrial Fibrillation Clopidogrel Trial with Irbesartan for Prevention of Vascular Events (ACTIVE-W), resulting in the halting of those trials.

Alternative approaches for stroke prevention in AF are sorely needed. For warfarin-ineligible patients, the combination of aspirin and clopidogrel remains under study in the ACTIVE-I trial. Because the left atrial appendage (LAA) is thought to be the primary source of emboli, endovascular closure of the LAA is under study in a randomized clinical trial comparing the Watchman device to adjusted-dose warfarin in warfarin-eligible patients. Endovascular ablation procedures have also been posed as an alternative, although one of the risks of the procedure is cardioembolic stroke, and the risk of AF relapse would still warrant long-term anticoagulation for stroke prevention. There are no pivotal randomized clinical trials comparing ablation to standard medical therapies, but aggregate case series have reported reductions in stroke (25).

In the Atrial Fibrillation Follow-up Investigation of Rhythm Management (AFFIRM) trial comparing a rate versus rhythm control strategy, all patients were initially anticoagulated, but this was discontinued in some patients who were believed to remain in sinus rhythm (26). The risk of stroke was related to the recurrence of AF as well as age, prior TIA or stroke, and diabetes but was not related to the rate or rhythm control strategy. Because warfarin therapy reduced stroke risk by 68%, the authors hypothesized that episodes of occult AF continue, and they recommended that anticoagulation be continued lifelong in eligible patients (27).

Prior to thrombolytic therapy, stroke complicated 0.8% to 5.5% of acute MIs. For most patients, the stroke risk continues for 4 to 6 months, although 90% occur within the first 2 weeks. Risk factors for stroke include older age, prior history of stroke, paroxysmal AF, anterior or apical location, impaired left ventricular (LV) function, and severity of MI. Mural thrombi have been identified in 38% to 67% of pathologic studies, but embolization often occurs in the absence of detectable thrombus on echocardiography. Three large randomized clinical trials have demonstrated stroke reduction from 2.3% to 5% to 0.8% to 1.7% with short-term anticoagulant therapy. Long-term anticoagulant therapy after acute MI has also been shown to reduce embolic stroke but with an increased risk for intracerebral hemorrhage (ICH); such therapy is recommended in patients at increased risk for embolic stroke such
as those with AF, prior systemic embolism, CHF, echo evidence of mural thrombus, or persistent LV dysfunction.

Embolism occurs with all degrees of rheumatic mitral valve disease but is more frequently a complication of mitral stenosis, in which stroke risk is increased 6-fold without and 18-fold with concomitant atrial fibrillation. Long-term anticoagulation is recommended with AF as well as mitral stenosis associated with enlargement of the left atrium greater than 5.5 cm, presence of left atrial spontaneous echo contrast, and prior to balloon valvuloplasty. The risk for prosthetic valve thromboembolism is higher with valves in the mitral than in the aortic position, with multiple than with single valves, and with caged-ball than with tilting-disc or bileaflet valves. Additional risk factors for thromboembolism include prior thromboembolism, AF, CAD, an enlarged left atrium, and left atrial thrombus. Recommendations for antithrombotic therapy mirror these risk factors.

Mitral valve prolapse was commonly invoked as a cause of cryptogenic stroke in young adults several decades ago. Although pathologic documentation of platelet-fibrin aggregates on the valvular surface confirm its causal potential, it is more likely to be the cause of stroke when other risk factors are present, such as AF, an atrial septal abnormality, or hypercoagulable state. Antiplatelet therapy with aspirin 160 to 325 mg/day is recommended for secondary stroke prevention, with anticoagulation reserved for those failing antiplatelet therapy.

Embolization of calcific debris can complicate valvuloplasty or replacement of calcific aortic stenosis, but spontaneous embolization is uncommon. Mitral annual calcification is identified in 20% to 30% of elderly patients and is more a marker of vascular risk factors than a source of embolism.

Greater than 90% of all cases of chronic heart failure are due to dilated cardiomyopathy, which encompasses nearly 50 distinct diseases. Anticoagulation with an INR range of 2.0 to 3.0 (target 2.5) is recommended when chronic heart failure becomes complicated by atrial fibrillation or pulmonary or systemic embolism. The risk of ischemic stroke correlates better with the ejection fraction (EF) than clinical symptoms and increases with worsening LV from 1.5%/year to 2%/year at EF less than 30% to 35% to 2%/year to 4%/year with EF less than 10%. In the absence of atrial fibrillation, the optimum antithrombotic therapy for patients with poor LV function remains controversial.

The VA Heart Failure Trial (WATCH) was prematurely terminated due to poor enrollment. There was no difference in the primary endpoint of stroke, MI, or death among those treated with warfarin (INR 2.5 to 3), aspirin 160 mg daily, or clopidogrel 75 mg daily, although warfarin-treated patients had less stroke but more bleeding complications, and aspirin-treated patients had more hospitalizations for heart failure. The ongoing NIH-funded Warfarin vs Aspirin in Reduced Cardiac Ejection Fraction (WARCEF) trial is studying warfarin versus aspirin 325 mg daily in 2,860 patients, with prespecified plans to combine the data with the WATCH results.

Atrial septal abnormalities have become increasingly recognized as a potential cause of cryptogenic stroke in young adults, particularly within the last decade with the development of endovascular closure devices for treatment. Although paradoxical embolism is implied, rarely is a thrombus found straddling the atria. Patent foramen ovale (PFO) is found with a greater frequency in young patients with cryptogenic cause (50% to 60%) than in those with known cause (30% to 50%) or controls (10% to 30%). The risk of recurrent ischemic stroke is low at 0% to 4% and may be increased by large PFO size, associated atrial septal aneurysm, mitral valve prolapse, associated atrial arrhythmias, or underlying coagulopathy. The inclusion of atrial septal aneurysm, a bulging of the interatrial septum identified on echocardiography, as a potential cardioembolic source of stroke stems from its identification in 10% to 20% of those with cryptogenic stroke compared to 1% to 4% of normal controls. Risk appears to be increased with large size (>1 cm of bulging) as well as associated abnormalities such as septal fenestrations, atrial septal defects, patent foramen ovale, and mitral valve prolapse.

Case series of endovascular PFO closure support the safety and feasibility of treatment, although recurrent ischemic events are not eliminated and no data are available demonstrating superiority to medical therapy with either aspirin or warfarin alone (28). Clinical trials are underway to determine whether endovascular PFO closure with septal repair implants offers superior protection against stroke and systemic embolism compared to medical therapy with warfarin or aspirin. Trials include the CLOSURE trial (Prospective, Multicenter, Randomized Controlled Trial to Evaluate the Safety and Efficacy of the STARFlex® Septal Closure System Versus Best Medical Therapy in Patients with a Stroke and/or Transient Ischemic Attack Due to Presumed Paradoxical Embolism through a Patent Foramen Ovale) of 1,600 patients using the STARFlex device (NMT Medical, Boston), the RESPECT (Randomized Evaluation of Recurrent Stroke Comparing PFO Closure to Established Current Standard of Care Treatment) trial of 500 patients using the Amplatzer PFO occluder (AGA Medical Corporation, Golden Valley, MN), and the CARDIA-Star trial of 300 patients using the CARDIA-Star PFO closure device (Cardia, Burnsville, MN). Because the effectiveness of these devices compared to standard medical therapies has not been demonstrated, endovascular closure outside of a clinical trial in the United States should be restricted to the clinical scenario outlined in the Food and Drug Administration (FDA) Human Device Exemption statement, which specifies patients with a recurrent cryptogenic stroke presumed to be due to paradoxical embolism failing therapeutic levels of anticoagulant therapy.

Hypercoagulable states are linked to stroke by several mechanisms. Inherited or acquired prothrombotic disorders such as antithrombin III deficiency, factor V Leiden, protein C deficiency, protein S deficiency, prothrombin gene 20210A mutation, paroxysmal nocturnal hemoglobinuria, and the nephrotic syndrome are more likely to produce venous thromboses. Cerebral venous thrombosis can produce venous infarction and cerebral hemorrhage, and systemic venous thrombosis can serve as a source of cerebral embolism if a concomitant intracardiac defect permitting right-to-left shunting is present (29).