Stroke is a common and devastating disease, the fifth leading cause of death and the leading cause of disability in the United States. On average, one American dies from a stroke every 4 minutes. Cardiogenic stroke can occur when (1) the heart pumps unwanted material into the circulation that reaches the brain (embolism), (2) pump function fails and the brain is hypoperfused, or (3) drugs given to treat cardiac disease have adverse neurologic effects.
Direct Cardiogenic Brain Embolism
Cardiogenic cerebral embolism is responsible for approximately 20% of ischemic strokes.1,2,3,4,5 Because many patients have coexisting cardiac and extracranial vascular disease,5 criteria for the diagnosis of cardiac embolism remain controversial even today. As more advanced diagnostic techniques have been developed, more causative cardiac abnormalities (and their association with stroke) have been recognized. Cardiac sources of brain emboli can be divided into several groups6:
Cardiac wall and chamber abnormalities
Aortic arch disease
Some cardiac sources have much higher rates of initial and recurrent embolism. The more common sources will be reviewed. The Stroke Data Bank7 in 1992 divided potential sources of brain embolism into strong sources (prosthetic valves, atrial fibrillation [AF], endocarditis, sick sinus syndrome, ventricular aneurysm, akinetic segments, mural thrombi, cardiomyopathy, and diffuse ventricular hypokinesia) and weak sources (myocardial infarct > 6 months old, aortic and mitral stenosis and regurgitation, congestive failure, mitral valve prolapse, mitral annulus calcification, and hypokinetic ventricular segments). Patients who have these weak sources are now often lumped within a category called cryptogenic stroke or cryptogenic embolism. The sources then deemed weak were frequent findings in patients who did not have brain embolism. Research is now directed into defining the frequency of these sources and identifying ways to determine in which patients they are the embolic source and in whom they are incidental findings. The risk of embolism varies within individual cardiac abnormalities depending on many factors. For example, in patients with AF, associated heart disease, patient age, duration, chronic versus intermittent fibrillation, and atrial size all influence embolic risk. The presence of a potential cardiac source of embolism does not mean that a stroke was caused by an embolus from the heart. Coexistent occlusive cerebrovascular disease is common. In the Lausanne Stroke Registry, among patients with potential cardiac embolic sources, 11% of patients had severe cervicocranial vascular occlusive disease (> 75% stenosis) and 40% had mild-to-moderate stenosis proximal to brain infarcts.5
Persistent and paroxysmal AF is a potent predictor of first and recurrent stroke, affecting more than 2.7 million Americans. The principal adverse consequence of AF is ischemic stroke, with AF patients being five times more likely to have a stroke than those without AF. In patients with brain emboli caused by a cardiac source, there is a history of nonvalvular AF in roughly one-half of all cases, of left ventricular (LV) thrombus in almost one-third, and of valvular heart disease in one-fourth.1,8 Stroke prevention in patients with AF and other heart diseases is discussed later in this chapter.
Patients with large anterior myocardial infarctions (MIs) associated with a LV ejection fraction < 40% and anteroapical wall-motion abnormalities are at increased risk for developing mural thrombus caused by stasis of blood in the ventricular cavity as well as endocardial injury with associated inflammation.2,8 Ventricular thrombi can also occur in patients with chronic ventricular dysfunction caused by coronary disease, hypertension, and dilated cardiomyopathy. Stroke is less common among uncomplicated MI patients but can occur in up to 12% of patients with acute MI complicated by a LV thrombus. The rate of stroke is higher in patients with anterior rather than inferior infarcts, and may reach up to 20% in those with large anteroseptal MI. The incidence of embolism is highest during the period of active thrombus formation during the first 1 to 3 months, with substantial risk remaining even beyond the acute phase in patients with persistent myocardial dysfunction, congestive heart failure, or AF.8,9
Current data indicate that congestive heart failure affects an estimated 5.1 million Americans. Patients with ischemic and nonischemic dilated cardiomyopathy have a similarly increased stroke risk by a factor of 2 to 3, accounting for an estimated 10% of ischemic strokes.3,8,10 Stroke rates may be higher in certain subgroups, including patients with prior stroke or trasient ischemic attack (TIA), lower ejection fraction, LV noncompaction, peripartum cardiomyopathy, and Chagas heart disease.8 The 5-year recurrent stroke rate in patients with cardiac failure has been reported to be as high as 45%.8,11
The magnitude of risk for brain embolism from a diseased heart valve depends on the nature and the severity of the disease. Atrial fibrillation often coexists with valve disease.
Mitral Stenosis/Rheumatic Mitral Valve Disease
Mitral stenosis is most commonly caused by rheumatic fever. The main proximate cause for embolic stroke in mitral stenosis of any cause is AF. Other factors associated with increased stroke risk in mitral stenosis patients include older age, left atrial enlargement, reduced cardiac output, and prior embolic event. In older studies prior to the era of chronic anticoagulation, recurrent embolism was reported in 30% to 60%.8,11,12,13,14 Between 60% and 65% of these recurrences developed the first year, many within the first 6 months.8,12,13 The majority of patients in these studies did have AF. Mitral valvuloplasty does not appear to eliminate the risk of embolism.8,15
Mitral valve prolapse (MVP) is the most common form of valve disease in adults and is generally benign.17,18 MVP as a source of embolic stroke continues to be controversial.6,19 Several small clinical series have reported cerebral embolism in MVP patients who lacked other possible embolic sources.19,20,21,22 In the absence of AF, MVP/mitral regurgitation is probably not associated with a significant increase in the risk of first or recurrent stroke.8,21,22
Mitral Annulus Calcification
Several series suggest a relation between mitral annulus calcification (MAC) and brain emboli and stroke.3,6,23,24,25 In the Framingham Heart Study, MAC was associated with an increased risk for all types of stroke over 8 years of observation; however, only 63% were embolic and a portion of those were also associated with AF. Other population-based studies did not reveal a significant association. The association between MAC and increased risk of stroke may be the result of shared risk factors rather than direct causation. For example, AF can occur 12 times more often in patients with MAC than in those without MAC.18,26
Neither aortic regurgitation nor aortic stenosis is associated with an increased risk of first or recurrent stroke. Additionally, studies of lesser degrees of aortic valve disease, including aortic annular calcification and aortic valve sclerosis, have also not confirmed an association with an increased risk of stroke.27,28
Stroke in many patients remains cryptogenic, and more patients may have cardiogenic embolism than are recognized. Clinical features and brain investigations such as computed tomography (CT), magnetic resonance imaging (MRI), and angiography (CT angiography [CTA], magnetic resonance [MR] angiography [MRA], and digital subtraction angiography) may suggest emboli, but often a clear source is unidentified. These cases, which are termed infarcts of unknown causes in the Stroke Data Bank,29,30,31 include as many as 40% of all stroke patients.
Fibrous and fibrinous lesions of the heart valves and endocardium are associated with certain medical conditions, with the specific stroke risk depending on the underlying medical condition.6 Valve lesions occur in patients with systemic lupus erythematosus (Libman-Sacks endocarditis32), antiphospholipid antibody syndrome,33 cancer, and other debilitating diseases (nonbacterial thrombotic endocarditis). Mobile fibrous strands are also often found during echocardiography.6,34,35,36 Fibrin-platelet aggregates may attach to these fibrous and fibrinous lesions.
Embolic complications are common in patients who have infective endocarditis.6,37 Mycotic aneurysms can cause fatal subarachnoid bleeding. Bleeding can also result from vascular necrosis as a result of an infected embolus.37 Embolization usually stops when infection is controlled.34 Warfarin does not prevent embolization and is contraindicated in patients with endocarditis and known cerebral embolism, unless there are other important lesions such as prosthetic valves or pulmonary embolism. In children and young adults with congenital heart defects, especially those with right-to-left shunts and polycythemia, brain abscess is an important complication.
Emboli often arise from sources other than the heart, such as the aorta, proximal arteries (intra-arterial or so-called local embolism), leg veins (paradoxical emboli), fat in the liver or bones (fat embolism), and materials introduced by the patient or physician (drug particles or air).6 The types of embolic materials vary (Table 94–1).6,38 Atheromatous plaques in the aortic arch and ascending aorta are an important source of embolism to the brain (Figs. 94–1 and 94–2). Ulcerated atheromatous plaques are often found at necropsy in patients with ischemic strokes, especially in those in whom the stroke etiology was not determined during life.39 Transesophageal echocardiography (TEE) often shows these atheromas, but technical factors limit visualization of the entire arch.40 Large (> 4 mm), protruding mobile aortic atheromas are especially likely to cause embolic strokes and are associated with a high rate of recurrent strokes.41,42 Use of oral anticoagulants rather than antiplatelet agents is recommended in these patients.18,43,44
TABLE 94–1.Embolic Materials ||Download (.pdf) TABLE 94–1. Embolic Materials
|Cardiac ||Intra-Arterial |
Red fibrin-dependent thrombi
White platelet-fibrin nidi
Material from marantic endocarditis
Bacteria from vegetations
Calcium from valves and mitral annulus calcification
Myxoma cells and debris
Red fibrin-dependent thrombi
White platelet-fibrin nidi
Combined fibrin-platelet and fibrin-dependent clots
Atheromatous plaque debris
Calcium from vascular calcifications
Mucin from tumors
Talc or microcrystalline cellulose from injected drugs
Descending aorta at necropsy from a patient whose transesophageal echocardiography before surgery showed severe disease of the ascending aorta and aortic arch with mobile protruding plaques. This patient died after coronary artery bypass grafting surgery having never awakened after the procedure. Used with permission from Denise Barbut, MD, Cornell University Medical College and the New York Hospital.
Cholesterol crystals and other particulate debris are caught in a filter placed in the aorta at the time that aortic clamps are removed. Used with permission from Denise Barbut, MD, Cornell University Medical College and the New York Hospital.
Clinical Onset and Course
Warning signs of stroke can include sudden hemiparesis, hemisensory loss, confusion, trouble speaking or understanding, visual loss, diplopia, ataxia, vertigo, or sudden severe headache with no known cause. Most embolic events occur during activities of daily living, but some embolic strokes have their onset during rest or sleep. Sudden coughing, sneezing, or nighttime micturation can precipitate embolism.6,45 Although the deficit is most often maximal at outset, 11% of embolic stroke patients in the Harvard Stroke Registry had a stuttering or stepwise course, whereas 10% had fluctuations or progressive deficits. Later progression, if it occurs, is usually within the first 48 hours. Progression is usually caused by distal passage of emboli. As in all large infarcts, brain edema and swelling may develop during the 24 to 72 hours after stroke with headache, decreased alertness, and worsening of neurologic signs. The edema is often cytotoxic (inside cells) and usually does not respond to corticosteroid treatment.
Emboli usually cause occlusion of distal branches and produce surface infarcts that are roughly triangular, with the apex of the triangle pointing inward. CT and MRI findings can suggest that an ischemic stroke was cardioembolic by the location and shape of the lesions on imaging46; eg, finding the presence of superficial wedge-shaped infarcts in multiple different vascular territories, hemorrhagic infarction, as well as visualization of thrombi within arteries. Among 60 patients with cardiogenic sources of embolism studied by CT in whom occlusive atherosclerotic cerebrovascular disease had been excluded, 56 had superficial large or small cortical or subcortical infarcts and only 4 had deep infarcts.46 Emboli can also block the middle cerebral artery leading to solely deep infarcts, as a result of collateral flow preservation of superficial cerebral arterial flow.6,45,47 Tiny emboli may also cause small deep or superficial infarcts.
MRI, particularly with the use of MR diffusion-weighted and MR gradient recall echo (GRE) imaging, is much more sensitive for detection of acute brain infarcts than is CT. MR is also superior in detecting hemorrhagic infarction by imaging hemosiderin. Hemorrhagic infarction has long been considered characteristic of embolism, especially when the artery leading to the infarct is patent.48 The mechanism of hemorrhagic infarction is reperfusion of ischemic zones after iatrogenic opening of an occluded artery (eg, endarterectomy, fibrinolytic treatment) or after restoration of the circulation after a period of systemic hypoperfusion. Hemorrhage then occurs into proximal reperfused regions of brain infarcts.6,45,49 At times, it is also possible to image the acute embolus on CT and also via MRI.6,50,51,52
In unselected series of stroke patients, transthoracic echocardiography (TTE) has been variably useful in detecting sources.6,53,54,55 TTE is useful in patients with known cardiac disease to clarify potential embolic sources and heart function,45 in young patients without stroke risk factors, and in stroke patients who do not have lacunar infarction or ultrasound or CTA evidence of intrinsic atherostenosis of a major extracranial and intracranial artery. TEE provides much better visualization of the aorta, atria, cardiac valves, and septal regions. Reports of TEE suggest that the diagnostic yield is 2 to 10 times that of TTE.56,57,58,59 Aortic plaques, atrial septal aneurysms, and atrial septal defects are also much better seen with TEE (Fig. 94–3). The use of an echo-enhancing agent such as agitated saline or echogenic contrast helps detect intracardiac shunts.
Transesophageal echocardiography recording during cardiac surgery from the aorta at the level of the origin of the left subclavian artery. A mobile plaque is seen protruding into the aortic lumen (small black arrow). This recording was taken after the release of aortic clamps and shows a shower of emboli within the aortic lumen beyond where the aorta was previously clamped. Used with permission from Denise Barbut, MD, Cornell University Medical College and the New York Hospital.
Echocardiography has definite limitations. Particles the size of 2 mm can block major brain arteries, but are beyond the imaging resolution of current echocardiographic technology.60 Also, thromboembolism is a dynamic process. When a clot forms in the heart and embolizes, there may be no residual evidence unless a clot reforms.6,38 Cardiac thrombi are imaged differently on sequential echocardiograms6,64; even large thrombi once seen on one echocardiogram can disappear later.61
Cerebral embolic signals can also be detected by monitoring with transcranial Doppler (TCD).6,62,63 Embolic particles passing under TCD probes produce transient, short-duration, high-intensity signals referred to as HITSs (high-intensity transient signals). Examples of HITSs are shown in Figs. 94–4 and 94–5. TCD monitoring of patients with AF,64 cardiac surgery,65 prosthetic valves, LV assist devices,66 carotid artery disease, during carotid endarterectomy, and in stroke patients with a patent foramen ovale have shown a relatively high frequency of embolic signals.67 Monitoring of emboli with TCD may help guide treatment decisions.
Transcranial Doppler recording from the middle cerebral arteries during steady-state cardiac bypass surgery at a time when the aorta was manipulated. The white streaks represent microemboli. Used with permission from Denise Barbut, MD, Cornell University Medical College and the New York Hospital.
Transcranial Doppler recording from the middle cerebral arteries during cardiac bypass surgery. A few distinct emboli (white streaks in the left of the figure) are followed by a massive shower of emboli (whiteout) at the time of the release of aortic clamps. Used with permission from Denise Barbut, MD, Cornell University Medical College and the New York Hospital.
Prevention and Treatment of Direct Cardiogenic Brain Embolism
Atrial Fibrillation: Treatment Options
Multiple studies have showed warfarin to be effective in preventing brain embolism in patients with both rheumatic mitral stenosis and AF. Anticoagulation reduces the risk of ischemic stroke (and other embolic events) by about two-thirds, irrespective of baseline risk. The intensity of anticoagulation was higher in early studies than that currently used, and brain hemorrhages and other bleeding complications were common. Trials have now shown that lower-dose warfarin (international normalized ratio [INR] 2.0-3.0) is effective in preventing brain emboli in patients with nonrheumatic AF.
In the Copenhagen Atrial Fibrillation, Aspirin, Anticoagulation (AFASAK) study, 1007 patients (median age, 74.2 years) with chronic, nonrheumatic AF were assigned to warfarin (INR 2.8-4.2), aspirin (75 mg/d), or placebo.68 The study was halted prematurely when analysis of effectiveness reached a predetermined level of significance in favor of warfarin treatment. The principal outcome was the composite of ischemic or hemorrhagic stroke, TIA, and systemic embolism. The observed reduction for warfarin compared with placebo was 64%, an absolute risk reduction of 3.5% per year. An analysis by intention to treat, which excluded TIA and minor stroke, indicated a risk reduction of approximately 50% (P < .05) and an absolute reduction of approximately 1.5% per year.
The Stroke Prevention in Atrial Fibrillation (SPAF) study investigators evaluated warfarin and aspirin in patients with nonrheumatic AF.69,70,71 The study evaluated two groups of patients based on their eligibility for warfarin. In the first group, 627 patients judged eligible for warfarin were randomized to open-label warfarin (INR 2.8-4.5; prothrombin time [PT], 1.3-1.8 times control) or, in a double-blinded fashion, to either aspirin (325 mg daily, enteric coated) or a matching placebo. In the second group, 703 patients considered ineligible for warfarin were randomized (double blind) to aspirin (325 mg daily, enteric coated) or placebo. The principal outcome, a composite of ischemic stroke and systemic embolism, was significantly decreased during a mean follow-up of 1.3 years. The outcome of disabling ischemic stroke or vascular death was reduced by warfarin by 54% (P = .11), an absolute reduction of 2.6% per year. The outcome of disabling stroke or death was reduced 22% by aspirin (P = .33), an absolute reduction of approximately 1% per year. The SPAF investigators later compared low-intensity, fixed-dose warfarin (INR 1.2-1.5) plus aspirin (325 mg/d) with adjusted-dose warfarin (INR 2.0-3.0) in elderly patients with one or more risk factors for embolism.71 Ischemic stroke and systemic embolism were present in 7.9% of patients on fixed-dose warfarin plus aspirin versus only 1.9% of patients on adjusted-dose warfarin. SPAF investigators later studied the effectiveness of aspirin 325 mg in patients with low risk and found that the rate of ischemic stroke was low (2% per year).72
The SPAF study identified three risk factors for thromboembolism: (1) recent congestive heart failure, (2) history of hypertension, and (3) previous thromboembolism.73,74 The study also suggested that anticoagulation with warfarin was not indicated in patients with none of the three risk factors who were at low risk for thromboembolism (2.5% per year). SPAF investigators concluded that in such patients, the dangers of anticoagulant therapy may outweigh its benefits. Aspirin (325 mg daily) is probably reasonable and safe therapy for patients with lone, nonrheumatic AF, who are younger than 60 years of age, and have none of the three identified risk factors.73,74,75 In all other patients with AF, long-term oral warfarin therapy (INR 2.0-3.0) is recommended unless contraindicated.72,75
In the Boston Area Anticoagulation Trial for Atrial Fibrillation (BAATAF), 420 patients with nonrheumatic AF (mean age, 68 years) were randomized unblinded to warfarin (target PT ratio, 1.2:1.5 × control; INR 1.5-2.7) or to a control group who were allowed to take aspirin.76 The principal outcome was ischemic stroke or systemic embolism, and the mean follow-up time was 2.2 years. The incidence of stroke was reduced by 86% in the warfarin group compared with the control group (P = .002), equivalent to an absolute risk reduction of 2.6% per year. There was no demonstrable benefit of aspirin, but the study was not designed to test aspirin.
Although many trials showed the superiority of adjusted-dose warfarin over antiplatelet therapy for stroke prevention, participants in those trials tended to be younger (typically 70 years old) than the AF patients now commonly encountered in clinical practice (typically late 70s with a substantial fraction of octogenarians). The Birmingham Atrial Fibrillation Treatment in the Aged (BAFTA) trial addressed this issue, randomizing 973 AF patients aged ≥ 75 years to adjusted-dose warfarin versus aspirin 75 mg/d. The stroke rate was 5% on aspirin and nearly halved by warfarin. Surprisingly, major hemorrhage rates were similar in both groups, but 40% of patients had received warfarin previously, potentially biasing toward lower bleeding rates. The investigators concluded that “these data lend support to the use of anticoagulation for all people aged over 75 years who have AF, unless there are contraindications or the patient decides that the size of the benefit is not worth the inconvenience of treatment.”77
Warfarin is approximately 50% more effective than aspirin in preventing stroke in patients with AF who do not have valvular disease. Data suggest that the optimal intensity of oral anticoagulation for stroke prevention in patients with AF appears to be a target INR of 2.0 to 3.0. However, the narrow therapeutic margin of warfarin, in addition to associated food and drug interactions, requires frequent INR testing and dosage adjustments. These liabilities likely contribute to underuse of warfarin, and alternative therapies are needed.
The Atrial Fibrillation Clopidogrel Trial with Irbesartan for Prevention of Vascular Events (ACTIVE) evaluated the safety and efficacy of the combination of aspirin plus clopidogrel in AF patients who were unsuitable candidates for vitamin K antagonist therapy. In those patients, the addition of clopidogrel to aspirin did reduce the risk of major vascular events, especially stroke, but also increased the risk of major hemorrhage.78
Novel Oral Anticoagulants (NOACs)
Oral direct thrombin inhibitors (DTIs): Dabigatran etexilate
Oral direct factor Xa inhibitors: Apixaban, rivaroxaban, otamixaban, betrixaban, and edoxaban
DTIs prevent thrombin from cleaving fibrinogen into fibrin. They bind to thrombin directly, rather than by enhancing the activity of antithrombin, as is done by heparin. The only oral DTI available for clinical use is dabigatran etexilate. Another oral agent, ximelagatran, was tested, but development was discontinued in 2006 because of hepatotoxicity and cardiovascular events.
Dabigatran has been compared with warfarin in the landmark RE-LY trial, which was an open, prospective, randomized study, with blinded adjudication of events. In this trial, 18,113 patients with AF were assigned to dabigatran 110 mg twice a day, dabigatran 150 mg twice a day, or adjusted-dose warfarin. There was a trend for less MI with warfarin. Compared to warfarin, dabigatran 110 mg twice a day had similar efficacy in terms of stroke rates, with fewer major bleeds and fewer hospitalizations. Dabigatran given at 150 mg twice a day had a lower stroke rate compared with warfarin, with less mortality but similar major bleeding rates. Of note, dyspepsia, dysmotility, and gastrointestinal reflux were twice as common in those who received dabigatran.79
Direct factor Xa inhibitors prevent factor Xa from cleaving prothrombin to thrombin. They bind directly to factor Xa, rather than enhancing the activity of antithrombin III. Examples of oral direct factor Xa inhibitors include apixaban, rivaroxaban, otamixaban, betrixaban, and edoxaban. The Rivaroxaban Once Daily Oral Direct Factor Xa Inhibition Compared with Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation (ROCKET AF) study was a prospective, randomized, double-blind, multicenter, event-driven, noninferiority study comparing the efficacy and safety of rivaroxaban, an oral, once-daily, direct factor Xa inhibitor, with adjusted-dose warfarin in patients with nonvalvular AF.80 Additionally, a different factor Xa inhibitor, apixaban, was also compared with warfarin in the ARISTOTLE trial.81
Ruff and colleagues (2014) performed the first meta-analysis of four novel oral anticoagulants: dabigatran, rivaroxaban, apixaban, and edoxaban.82 Anticoagulation with each of these NOACs led to similar or lower rates both of ischemic stroke and major bleeding compared to adjusted-dose warfarin (INR 2.0-3.0) in patients with nonvalvular AF. Important additional advantages of the NOAC agents are convenience (no requirement for routine testing of the INR), a high relative but small absolute reduction in the risk of intracerebral hemorrhage (ICH), lack of susceptibility to dietary interactions, and reduced susceptibility to drug interactions. Disadvantages include lack of efficacy and safety data in patients with chronic severe kidney disease, lack of easily available monitoring of blood levels and compliance, higher cost, variable availability of agents that can reverse NOAC anticoagulation, as well as the potential for unanticipated side effects, which will subsequently become evident.79,80,81,82
Additional meta-analyses have pooled the results from the RE-LY, ARISTOTLE, ROCKET AF, and ENGAGE AF-TIMI 48 trial and come to similar conclusions.79,80,81,82,83,84,85,86 More recently, a 2014 Cochrane review compared the following factor Xa inhibitors to warfarin in patients with AF: apixaban, betrixaban, darexaban, edoxaban, idraparinux, and rivaroxaban. A lower rate of stroke and systemic embolic events was found with the NOACs, as well a lower rate of death and ICH.87 A second 2014 Cochrane review evaluated studies that compared direct thrombin inhibitors to warfarin and found no significant difference in the odds of vascular death and ischemic events.88 Fatal and nonfatal major bleeding events, including hemorrhagic strokes, were less frequent with these agents.
These meta-analyses support the broad concept that NOAC agents (direct thrombin and factor Xa inhibitors) are preferable to warfarin use in patients with AF. However, they cannot directly compare the relative advantages and disadvantages of the individual agents nor can they demonstrate whether the agents are equal in safety and efficacy. Also, the NOACs were studied in nonvalvular AF, and as a result, they are only approved for use in nonvalvular AF.
There is no single perfect medication choice for patients with AF. The authors assert that a variety of factors should influence a physician’s selection of therapy for their patient. Assessing a particular patient’s risk of ischemic stroke with CHA2DS-VASc scoring can assist in making a treatment decision between antiplatelet, warfarin, and a NOAC.89 For example, current 2014 American College of Cardiology (ACC)/American Heart Association (AHA)/Heart Rhythm Society (HRS) guidelines note that for patients with a CHA2DS2-VASc score of 0, antithrombotic therapy may be omitted (class IIa, level of evidence B).90 Additionally, assessing the patient’s risk of significant bleeding with HAS-BLED scoring can also influence medication choice.91 We recommend that therapy should be tailored to the individual patient and their specific needs, including making an assessment of other comorbid medical issues, cost of the particular agent selected, history of patient compliance with medications, as well as patient preference.
The effectiveness of anticoagulation on embolic stroke prevention from other cardiac conditions has not been well studied. Current AHA/American Stroke Association (ASA) guidelines recommend the following in patients with ischemic stroke or TIA in the setting of MI with LV thrombus or apical wall-motion abnormality8:
Treatment with warfarin (target INR, 2.5; range, 2.0-3.0) for 3 months is recommended in most patients with ischemic stroke or TIA in the setting of acute MI complicated by LV mural thrombus formation identified by echocardiography or another imaging modality (class I; level of evidence C). Additional antiplatelet therapy for cardiac protection may be guided by recommendations such as those from the American College of Chest Physicians (ACCP).
Treatment with warfarin (target INR, 2.5; range, 2.0-3.0) for 3 months may be considered in patients with ischemic stroke or TIA in the setting of acute anterior ST-segment elevation MI (STEMI) without demonstrable LV mural thrombus formation but with anterior apical akinesis or dyskinesis identified by echocardiography or other imaging modality (class IIb; level of evidence C).
In patients with ischemic stroke or TIA in the setting of acute MI complicated by LV mural thrombus formation or anterior or apical wall-motion abnormalities with an LV ejection fraction < 40% who are intolerant to vitamin K antagonist (VKA) therapy because of nonhemorrhagic adverse events, treatment witha low molecular weight heparin (LMWH), dabigatran, rivaroxaban, or apixaban for 3 months may be considered as an alternative to VKA therapy for prevention of recurrent stroke or TIA (class IIb; level of evidence C).
Warfarin may not be effective in preventing calcific, myxomatous, bacterial, and fibrin-platelet emboli, and warfarin has been posited to worsen cholesterol crystal embolization.92
Timing of Warfarin Initiation After Embolic Stroke
Additionally the timing of the initiation of warfarin anticoagulation after embolic stroke remains controversial. Embolic brain infarcts often become hemorrhagic, and serious brain hemorrhage has occurred after anticoagulation.93,94,95 Large infarcts, hypertension, large bolus doses of heparin, and excessive anticoagulation have been associated with hemorrhage. Because most hemorrhagic transformations occur within 48 hours, the recommendations of the Cerebral Embolism Task Force were to avoid early anticoagulation in patients with large infarcts or hemorrhagic transformation on repeat CT.96,97 Studies of patients with cerebral and cerebellar hemorrhagic infarction show that, in the vast majority, the cause is embolic, with hemorrhagic infarction occuring equally with and without anticoagulation. The development of hemorrhagic infarction is rarely accompanied by clinical worsening.98,99 Patients with hemorrhagic transformation who were continued on anticoagulants in general did not worsen. The risk of re-embolism must be balanced against the small, but definite, risk of important bleeding. However, if the patient has a large brain infarct, heparin should be delayed, and bolus heparin infusions should be avoided. If the risk for re-embolism is high, immediate heparinization is advisable; whereas if the risk seems low, it is prudent to delay anticoagulants for at least 48 hours. One study showed that patients with AF with embolic strokes who were treated with well-controlled heparin anticoagulation soon after stroke onset fared better than patients treated later.100,101
Cryptogenic strokes, or strokes without a clear determined cause, remain a major challenge. Paradoxical embolism has been strongly implicated as one potential cause of cryptogenic stroke. Although once considered rare, emboli entering the systemic circulation through right-to-left shunting of blood are now often recognized with the advent of newer diagnostic technologies. By far the most common potential intracardiac shunt is a residual patent foramen ovale (PFO). The high frequency of PFOs in the normal adult population has made it difficult to be certain in an individual stroke patient with a PFO whether paradoxical embolism through the PFO was the cause of the stroke or whether the PFO was merely an incidental finding. Autopsy series have shown that approximately 30% of adults have a probe PFO at necropsy.102 Echocardiographic studies have shown that PFOs are more common in patients with an undetermined cause of stroke than in those in whom another etiology has been defined.103,104 Lechat and coworkers,103 using TTE with contrast injection during Valsalva maneuver, demonstrated right-to-left shunting through a PFO in 56% of patients with cryptogenic stroke, in comparison to 10% of the patients in the control group. Webster and colleagues,105 in a study of stroke patients younger than 40 years of age, found a PFO in 50% of patients with stroke using contrast echocardiography.
Neuroimaging studies are not conclusive with regard to the link between PFO and embolic stroke. However, in 1998, Steiner and colleagues106 reported on a series of 95 patients with first stroke and who had PFOs. Those with large PFOs had more features of embolic strokes with brain imaging than did patients with small PFOs.
Review of series of patients with paradoxical embolism107,108,109 through a PFO and the authors’ experience allow the derivation of five criteria that, when four or more are met, establish the presence of paradoxical embolism with a high degree of certainty6:
Situations that promote thrombosis of the deep veins of the leg or pelvis (eg, long sitting in one position, such as prolonged airplane flight, or recent surgery).
Increased coagulability (eg, the use of oral contraceptives, presence of factor V Leiden, dehydration, and other inherited or acquired thrombophilia).
The sudden onset of stroke during Valsalva or other maneuvers that promote right-to-left shunting of blood (eg, sexual intercourse, straining at stool).
Pulmonary embolism within a short time before or after the neurologic ischemic event.
The absence of other putative causes of stroke after thorough evaluation.
But the question remains: While a PFO may have contributed to the development of an ischemic stroke, is an isolated PFO capable of causing the stroke alone? The presence of an atrial septal aneurysm along with a PFO increases the risk of stroke. Among 581 patients the stroke risk was 2.3% in patients with PFO and 15.2 % in those with both PFO and atrial septal aneyrysm.109a The Risk of Paradoxical Embolism (RoPE) study performed a patient-level meta-analysis of 12 cryptogenic stroke cohorts.110,111 Among 3023 patients with cryptogenic stroke, the prevalence of PFO, and the likelihood that PFO was the cause of the stroke (the PFO-attributable fraction), correlated with the absence of vascular risk factors (ie, hypertension, diabetes, smoking, prior stroke or TIA, older age) and the presence of a cortical (as opposed to subcortical) cryptogenic infarct on imaging.110 Using multivariate modeling, the investigators devised the RoPE score, which estimates the probability that a PFO is either incidental or pathogenic in a patient with cryptogenic stroke.110 High RoPE scores, as found in younger patients who lack vascular risk factors and have a cortical infarct on neuroimaging, suggest pathogenic PFOs, while low RoPE scores, as found in older patients with vascular risk factors, suggest incidental PFOs. For each RoPE score stratum, the corresponding PFO prevalence was used to estimate the PFO-attributable fraction: the probability that the index event was related to the PFO. In a subsequent analysis, stroke recurrence was associated with the following three variables only in the high RoPE score group: a history of prior stroke or TIA, a hypermobile interatrial septum (atrial septal aneurysm), and a small shunt.111 The RoPE data did not include activity at onset; strokes that develop suddenly during sex, straining at stool, and Valsalva maneuvers and sudden exertion are often embolic.
In the author’s opinion, the following evaluation should be done in patients with acute stroke who are also found to have PFO to attempt to ensure there are no other causes for the stroke:
Brain imaging: MRI of brain without contrast if possible, CT of head without contrast if obtaining an MRI is prohibitory.
Vascular imaging: CTA of head and neck or MRA of head and neck or Duplex of the carotid arteries and TCD of head.
TEE with bubble study.
Hypercoagulable work-up: protein C, protein S, activated protein C resistance (if positive, test for factor V Leiden), dilute Russell viper venom time, cardiolipin antibody (if positive, test beta-2 glycoprotein 1 antibody), prothrombin gene mutation, antithrombin III, homocysteine
Other laboratory work-up: cholesterol panel, HBA1C, PT, INR, partial thromboplastin time, complete blood count, Chem7, troponin × 3, erythrocyte sedimentation rate. If patient is febrile or there is risk for endocarditis: blood cultures × 3.
Holter monitor, and if negative, longer monitoring as an outpatient for occult arrhythmia.
Current treatment options for future stroke prevention in patients with PFO and cryptogenic ischemic stroke include medical therapy, open or minimally invasive cardiac surgical closure, and transcatheter closure. Before any treatment option is selected, it is again important to confirm that the stroke is indeed cryptogenic. The treating physician should exclude all other possible contributing causes to the stroke, including a coexisting hypercoagulable state and deep venous thrombosis (with or without May-Thurner syndrome).113,114 If there is concomitant hypercoagulability or deep venous thrombosis, anticoagulation is the treatment of choice for these conditions as well as to prevent further ischemic stroke.
Once this has been done, with regard to medical therapies, antiplatelet therapy has always been considered reasonable for future stroke prevention in cryptogenic stroke patients with a first ischemic stroke/TIA plus an isolated PFO. In patients with a cryptogenic stroke and an atrial septal aneurysm, evidence had been insufficient to determine whether warfarin or aspirin is superior in preventing recurrent stroke or death. Warfarin is considered to be an appropriate treatment option in the subgroup of PFO/ischemic stroke patients with concomitant hypercoagulable state or venous thrombosis.
An alternative option is transcatheter PFO closure, a minimally invasive endovascular procedure during which a closure device is guided utilizing catheters to seal the PFO. Recently, transcatheter closure demonstrated a benefit compared to medical therapy for future stroke prevention. The RESPECT trial was a multicenter, prospective, randomized clinical trial designed to investigate whether percutaneous PFO closure, using the AMPLATZER PFO Occluder, is superior to current standard of care medical treatment in the prevention of recurrent embolic stroke. Investigators enrolled 980 patients (aged 18 to 60 years) with a PFO who experienced a cryptogenic stroke in the preceding 270 days. Patients were randomly assigned to receive either a PFO occluder (Amplatzer, AGA Medical, Golden Valley, MN; n = 499) or guideline-directed medication (n = 481). RESPECT represents the largest randomized trial on PFO closure ever conducted, with the longest-term follow-up, with a mean of 5 years and a duration of more than 10 years.115
Initial intent-to-treat analysis results revealed no significant difference in all-cause stroke between groups (P = .16), similar to prior PFO closure device studies. However, when strokes were restricted to cryptogenic strokes alone, Carroll and colleagues reported a 54% relative risk reduction in the PFO occlusion arm (P = .042), although the number of these strokes remained small (10 vs 19). Among patients under the age of 60 years, researchers found a 52% relative risk reduction in the PFO closure arm (P = .035). Furthermore, a 75% relative risk reduction in cryptogenic stroke was observed among PFO closure patients who had PFO characteristics of substantial shunt or atrial septal aneurysm (P = .007). In safety analysis, there were no cases of device erosion, embolization, or thrombosis, and no intra-procedure strokes. Additionally, the rate of AF was not significantly different between groups. Extended follow-up data from the RESPECT trial suggests that compared with medical management, a PFO closure device significantly decreases recurrent cryptogenic ischemic stroke in patients with PFO who had a cryptogenic stroke in the last 270 days.115
With regard to surgical closure of symptomatic PFO, there is no clear evidence at present that it is superior to medical or endovascular therapy for secondary stroke prevention. More recently, widespread use of intraoperative TEE during cardiac surgery has resulted in frequent discoveries of incidental asymptomatic PFOs. A recent survey by Sukernik and associates116 suggests that a number of cardiac surgeons in the United States alter their planned procedure to include closure of the PFO. This is concerning in light of recent data published by Krasuski and coworkers,117 who investigated the prevalence of intraoperative PFO diagnosis and the relationship that the incidental repair had on perioperative outcome and long-term survival. Surgeons were more likely to repair PFOs in patients who were younger, female, undergoing tricuspid or mitral surgery, had left atrial dilation, or with prior stroke and TIAs. Repair also tended to occur in patients with fewer comorbidities. Patients with incidental PFO were no more likely to have had a preprocedural stroke than patients without PFO. Postoperatively, the patients with an incidentally repaired PFO had 2.47 times greater odds of having an in-hospital, symptomatic postoperative stroke compared with those with unrepaired PFO (there was no difference in long-term survival).117 These findings of increased short-term postoperative stroke risk should discourage the routine closure of incidentally detected PFO for now. Further studies to assess whether any subgroup of PFO patients may benefit from closure will be important in the future.
Brain Hypoperfusion (Cardiac Pump Failure)
After cardiopulmonary resuscitation (CPR), the heart often recovers in individuals whose brain has been irreversibly damaged by ischemic-anoxic damage.118 Cardiologists must become very familiar with the pathology, signs, and prognosis of brain dysfunction after periods of circulatory failure.
Different brain regions have selective vulnerability to hypoxic-ischemic damage. Regions that are most remote and at the edges of major vascular supply are more liable to sustain hypoperfusion injury. These zones are usually referred to as border zones or watersheds. The cerebral cortex and hippocampus are particularly vulnerable to injury.119,120,121,122 In the cerebral cortex, the border zone regions are between the anterior cerebral artery (ACA) and middle cerebral artery (MCA), and between the MCA and posterior cerebral artery (PCA). The basal ganglia and thalamus are most involved if hypoxia is severe, but some circulation is preserved. This situation applies most to hanging, strangulation, drowning, and carbon monoxide exposure.123 Cerebellar neurons may also be selectively injured.124
When circulatory arrest is complete and abrupt, brainstem nuclei are especially vulnerable to necrosis in young humans and experimental animals.125 When hypoxia and ischemia are especially severe, the spinal cord may also be damaged.126,127 When cortical damage is severe and protracted, cytotoxic edema causes massive brain swelling, with cessation of blood flow and brain death.
Very severe hypoxic-ischemic damage can lead to mortal injury to the cortex and brainstem, irreversible coma, and brain death. When initially examined, such patients have no brainstem reflexes and no response to stimuli, except perhaps a decerebration response. These findings do not improve, and respiratory control is absent or lost.
When cerebral cortical damage is very severe, but brainstem reflexes are preserved, there is no meaningful response to the environment. Automatic facial movements such as blinking, tongue protrusion, and yawning usually persist. The eyes may rest slightly up and move from side to side. When this state does not improve, it is referred to as the persistent vegetative state118,121,128,129 or wakefulness without awareness. Laminar cortical necrosis can cause seizures (multifocal myoclonic twitches or jerks of the facial and limb muscles), which are difficult to control with anticonvulsants.
With severe hypoperfusion ACA-MCA border zone injury, there is weakness of the arms and proximal lower extremities with preservation of face, leg, and foot movement (the “man in a barrel” syndrome). With MCA-PCA ischemia, the symptoms and signs are predominantly visual. Patients describe difficulty seeing and inability to integrate the features of large objects or scenes despite retained capacity to see small objects in some parts of their visual fields. Reading is impossible. There are features of Balint syndrome.118,130 Apathy, inertia, and amnesia are also common. Patients cannot make new memories and have patchy, retrograde amnesia for events during and before hospitalization. This Korsakoff-type syndrome is caused by hippocampal damage and may not be fully reversible. Amnesia may be accompanied by visual abnormalities, apathy, and confusion, or may be isolated.
Shortly after resuscitation or arrest, patients with less severe cerebral injuries show some reactivity to the environment. Eye opening and restless limb movements develop. The eyes may fixate on objects. Noise, a flashlight, or a gentle pinch may arouse patients to react to stimuli. Soon patients awaken fully and may begin to speak. Cognitive and behavioral abnormalities may be detected after the patient awakens, depending on the degree of injury.
Prognostic signs and variables have been extensively studied.118,131,132,133,134 The initial neurologic findings and their course are helpful in predicting outcome. Among patients who have meaningful responses to pain at 1 hour, almost all survivors have preserved intellectual function. Patients who do not respond to pain by 24 hours typically either die or remain in a vegetative state. Being comatose predicts a poor prognosis.133,134 Two simple observations—the presence or absence of coma and the response to pain—predict neurologic outcome very early.134 Recurrent myoclonus is also a poor prognostic sign.135
In a study in Seattle of out-of-hospital cardiac arrests, patients who did not awaken died on average 3.5 days after arrest.136,137 Of 459 patients, 183 never awakened (40%). Among those who did awaken, 91 (33%) had persistent neurologic deficits.136 Prognosis could be made by analysis of pupillary light reflexes, eye movements, and motor responses.137 It is unclear whether bystander initiation of CPR is significantly related to awakening.137,138 After in-hospital CPR, pneumonia, hypotension, renal failure, cancer, and a housebound state before hospitalization were significantly related to death in the hospital.139
Neuroimaging and other tests have proved to be relatively unhelpful in contrast to the neurologic examination.118 CT is used to exclude other causes of coma, such as brain hemorrhage. Electroencephalography is helpful in studying cortical activity in unresponsive patients. TCD may be helpful in the evaluation of brain death, but is not a requirement to determine brain death.140,141,142
Other than maintaining adequate circulation and oxygenation, treatment has not helped improve outcome. Increased blood sugar correlates with poor outcome.143 A multifaceted approach to therapy has been most successful.144
Neurologic Effects of Cardiac Drugs and Cardiac Encephalopathy
Drugs given to patients with cardiac disease often have neurologic adverse effects.145 Digoxin can cause visual hallucinations, yellow vision, and general confusion.146,147 Digoxin levels need not be excessively elevated; the symptoms disappear with drug cessation. Quinidine can cause delirium, seizures, coma, vertigo, tinnitus, and visual blurring.148 Similar toxicity has been seen with lithium. Patients may become acutely comatose while being treated with intravenous lidocaine. This effect has been associated with the accidental administration of very large doses; more common CNS effects of less extreme toxicity include sedation, irritability, and twitching. The latter may progress to seizures accompanied by respiratory depression. Amiodarone can cause ataxia, weakness, tremors, paresthesias, visual symptoms, a Parkinsonian-like syndrome, and occasionally delirium.145 The neurological side effects with amiodarone can occur even at normal doses.
Patients with congestive heart failure often develop an encephalopathy characterized by decreased alertness, sleepiness, decrease in all intellectual functions, asterixis, and variability of alertness and cognitive functions from hour to hour.145 These patients may not have pulmonary, liver, or renal failure or electrolyte abnormalities. This cardiac encephalopathy is probably multifactoral.145