Standard 12-lead ECG is mandatory in all patients with acute stroke. ECG abnormalities are observed in 60–90 % of the patients with acute stroke [7, 8]. Not only heart rhythm disorder like atrial fibrillation (AF) but also diverse ECG changes following acute stroke can be detected on ECG. Myocardial ischemia-like ECG changes, such as ST-segment deviation, T wave abnormalities, and QTc prolongations are most frequent secondary ECG changes after acute stroke, which can mimic acute myocardial ischemia.

AF detection rate in acute stroke patients varies from 7 to 25 % [9–11]. Absence of AF in standard ECG test cannot exclude AF as a cause of the stroke event. Dedicated effort to detect paroxysmal AF possibly affect the AF detection rate. One study demonstrated that serial ECG follow-up during 72 h from stroke event can increase the chance of AF detection by 2.6-fold [12]. Continuous cardiac monitoring for at least 24 h is reasonable to detect AF and other serious arrhythmias [13, 14].

Arrhythmias other than AF also can be associated with acute phase of stroke. Serious arrhythmias, such as ventricular tachycardia, supraventricular tachycardia, various degrees of atrioventricular block, as well as AF, can be seen in approximately 25 % of the patients with acute stroke in first 3 days [15]. Arrhythmic events are especially frequent in the first 24 h and decline with time.

There is no strong recommendation of patient selection, time, and duration of Holter monitoring in patients with acute stroke. Detection rate of paroxysmal AF on 24-h Holter monitoring after ischemic stroke is between 2 and 7 % [16–18]. Some limited studies expressed a doubt about the efficacy of routine Holter monitoring to detect AF and other serious arrhythmias after acute stroke, because of the low detect rate and poor cost-effectiveness [16, 17, 19]. However, such findings might result from unselected patients, varying time and duration of monitoring. The Holter monitor can be especially effective in detecting cardiac rhythm disorder in patients with embolic infarction pattern, old age, and concomitant coronary artery disease [20]. Stroke patients with large deficits or right hemispheric stoke are at risk of various cardiac pathologic conditions, such as AF, myocardial ischemia, congestive heart failure, and other serious cardiac arrhythmias, and may need close cardiac monitoring and Holter monitoring [21]. One study suggested that routine Holter monitoring for over 24 h can identify new AF/atrial flutter in 1 out of 20 patients [18].

The longer duration of ECG monitor definitely increases the detection rate of AF and other cardiac arrhythmia. Table 6.2 demonstrated monitoring type and detection of paroxysmal AF [22]. Seven days of ECG monitoring using external event recorder revealed 16–25 % patients with paroxysmal AF among the patients who did not show AF in 24-h Holter monitoring [23, 24]. More prolonged monitoring may be considered up to a year or more using implantable loop recorder in patients with cryptogenic stroke or highly suspected heart rhythm disorder like AF. One study including patients with cryptogenic stroke revealed that 16 % of the patients had an event of paroxysmal AF using an implantable loop recorder during over a year of follow-up [25].

Echocardiography can provide information about functional and structural abnormalities inside and even outside of the heart. We should focus on the size and functions of left atrium (LA) and left atrial appendage (LAA), right ventricular and left ventricular (LV) function and regional wall motion abnormality, valvular disease, prosthetic valve, intracardiac shunt, and atheromatous embolic sources of aorta. All ischemic stroke patients with embolic character or without other clear etiology may need to undergo TTE. Although, some data suggested that transesophageal echocardiography (TEE) presented higher diagnostic yields in detecting embolic source compared to TTE [26, 27], TTE still has strong clinical advantages, which is noninvasive test, relatively lower price, and easier to perform and repeat. Furthermore, TTE is often better to evaluate native or prosthetic valve function and usually better for the detection of LV thrombus compared to TEE.

TTE is presenting very high sensitivity and specificity for detecting LV thrombus (Fig. 6.1) [28]. It is also an essential part to identify the presence of regional or global wall motion abnormality of left ventricle related to LV thrombus. However, in some patients who have poor echocardiographic window or spontaneous echocontrast around the suspected area, it is quite challenging to define thrombus. Contrast echocardiography technique can help to visualize thrombus in such cases. Computed tomography and cardiac magnetic resonance imaging can be options otherwise.

In patients with AF, most common site of the thrombus formation is LAA. TTE has limitations on visualizing LAA thrombus in patients with AF compared with TEE. LAA emptying velocity, which can be measured by LAA contraction using TEE by pulsed Doppler, is a parameter of the LAA dysfunction, and low velocity (≤20 cm/s) is related to risk of thrombus formation [29]. Some data showed that new technique of second harmonic imaging allows TTE to evaluate LAA function with better sensitivity [30, 31].

One of the most serious conditions causing embolic event is infective endocarditis (IE). Clinically, diagnosis of IE is made on basis of clinical, echocardiographic, and biological findings. Modified Duke criteria have been widely used in diagnosis of IE (Table 6.3) [32]. Vegetation, abscess, valvular regurgitations, and prosthetic valve dehiscence are frequently observed in IE (Fig. 6.2). Echocardiographic findings play a key role in diagnosis of IE. In patients with IE, embolic risk can be estimated by echocardiographic findings, such as size and mobility of the vegetation. Vegetation size in response to antibiotic therapy is also associated with risk of embolic event [33–36]. Vegetation size >10 mm tends to be highly embolic [37].

Paradoxical embolism is originated from venous side and cross to the systemic circulation through a shunt. Intracardiac shunt could be detected by echocardiography. TTE usually detect secundum- or primum-type atrial septal defect (ASD). Chamber size and pulmonary artery pressure also should be evaluated for further management. TEE is often needed in case of sinus venosus defect. Patent foramen ovale (PFO) also can cause paradoxical embolisms. Traditionally, TEE with Valsalva maneuver after agitated saline injection is the gold standard test for detecting PFO. However, with recent advance in techniques, TTE may be sufficient for the agitated saline test by using second harmonic imaging [38, 39].

Cardiac tumor frequently causes embolic event [40]. TTE provides not only anatomical information but also functional significance in those cardiac tumors. Cardiac myxoma is the most common primary tumor in the heart. Approximately 90% of cardiac myxoma occurred in LA [41], which can be highly associated with cardioembolic stroke (Fig. 6.3). Most common valve-associated tumor, fibroelastoma, is also related to embolic risk [42, 43]. Differential diagnosis between fibroelastoma and Lambl’s excrescences or infective endocarditis is often challenging.

Although TEE is an invasive, more expensive procedure which needs experienced physician, complication rate is less than 0.02 % [44], and diagnostic sensitivity can be superior than that of TTE especially for detection of thromboembolic source [26]. TEE can provide better spatial resolution of intracardiac structure even in patients with poor transthoracic window. Because TTE and TEE have their own pros and cons, respectively, the physician should make proper decision based on clinical circumstances as well as the cost-effectiveness.

TEE can sensitively detect LAA thrombus directly and highly thrombogenic conditions like spontaneous echocontrast in patients with AF. Doppler measurement in LAA is also a useful tool for evaluating thrombogenic risk in atrial fibrillation. LAA is not an immobile structure but is believed to have contraction and dynamic flow changes inside of it. LAA contraction forced blood out, and LAA emptying velocity is considered as a parameter of LAA function. The velocity below 20 cm/s is associated with risk of thrombus formation in LAA [45]. LAA emptying velocity also can be a parameter for successful cardioversion in patients with AF.

Regarding paradoxical embolism, sinus venosus-type or coronary sinus-type ASDs are often difficult to be found in TTE. TEE provides higher resolution around the inferior and superior vena cava area and even in the coronary sinus area. TEE also is a gold standard for PFO diagnosis. In differential diagnosis between PFO and pulmonary arteriovenous malformation as a source of paradoxical emboli, TEE test can be more accurate.

Presence of prosthetic valve is a highly thrombogenic condition. Although TTE has an important role in evaluating prosthetic valve function, TEE is required in many cases. In circumstances when prosthetic valve malfunction is present, differential diagnosis between pannus formation and thrombus formation is a crucial part in therapeutic strategy. High-resolution TEE image can provide important additional information in differential diagnosis [46].

Approximately 60 % of the patients 60 years or older who had stroke have aortic arch atherosclerosis (Fig. 6.4) [47]. Complex atheroma with ulceration and high mobility is closely related to the embolic stroke. TTE can approach aortic arch in suprasternal window, but imaging quality and information are often limited. TEE provides high-resolution images of ascending, descending aorta, and aortic arch. Thickness of atheroma and ulceration and presence of highly mobile portions can be assessed to estimate the embolic risk [45].

Dilatation and contractile dysfunction of the left atrium (LA) and left atrial appendage (LAA) in AF are significantly associated with blood stasis and thrombus formation. Ischemic stroke patient with AF can present worse short-term and long-term survival than those without, as well as higher recurrence rate of stroke [9]. There is a large body of evidence that anticoagulation is recommended in patients with AF according to their risk factors. Risk stratification using CHA2DS2-VASc score has been widely used to determine anticoagulation (Tables 6.4 and 6.5) [48]. Many studies showed that benefit of anticoagulation overweighs the risk in patients with AF and elevated risk of embolization. Thus, all patients with stroke/TIA who have AF got at least 2 points and should be anticoagulated unless there is contraindication for anticoagulation.

Traditionally, anticoagulation using vitamin K antagonist, warfarin, had been widely used in stroke prevention of AF patients. Recently, new anticoagulation agent blocking a specific target of coagulation cascade became popular, as an alternative to warfarin. Dabigatran, rivaroxaban, apixaban, and edoxaban are non-vitamin K antagonist oral anticoagulant, abbreviated as NOAC. NOACs demonstrated non-inferior efficacy in prevention of ischemic stroke compared to warfarin and secured their safety profiles in patient with non-valvular AF [49–52]. Because NOACs have their own pharmacokinetics, different renal excretion, side effect, and drug interaction, selection of the agent should be based on patients’ condition, risk factors, economic status, tolerability, and preference.

For patients who are not indicated for or cannot tolerate anticoagulation, antiplatelet treatment may be considered. Options can be aspirin monotherapy or dual antiplatelet therapy in combination with clopidogrel. However, evidences are not supporting the efficacy of aspirin preventing stroke in patients with AF [53]. Dual antiplatelet therapy in combination of aspirin and clopidogrel may be superior to aspirin monotherapy in preventing stroke but at the cost of increased bleeding [49].

Incidence of LV thrombus within 3 months after acute myocardial infarction (MI) vary from 3 to 16 % [54–57]. Recently, more aggressive antiplatelet therapy in acute stage may reduce the incidence of thrombus formation after MI. More than 10 % of the patients with LV thrombus are at risk of embolic event unless they are treated with anticoagulation [58, 59]. Available data is supporting the use of warfarin to reduce embolic risk at the presence of LV thrombus after MI. One meta-analysis data demonstrated 86 % reduction of embolic risk with warfarin in patients with LV thrombus after anterior MI [59]. It is reasonable to start unfractionated heparin or a low molecular weight heparin as soon as possible before warfarin effect gets to the therapeutic rage of international normalized ratio (INR) 2–3. There is limited data on the duration of anticoagulation therapy. Decision may need physician’s discretion based on the degrees of LV dysfunction, LV reverse remodeling during follow-up, and recurrence of thrombi. There is also limited data available about NOAC reducing the embolic risk of patients with LV thrombus.

Patients with mitral stenosis (MS) who suffered from ischemic stroke are recommended to receive warfarin therapy regardless of accompanying AF. In mitral stenosis patients with sinus rhythm who had ischemic stroke, paroxysmal AF or complicated infective endocarditis should be considered as an aggravating factor for stroke. Limited data suggested that adding aspirin to warfarin can reduce major adverse event compared to warfarin alone [60]. Regarding NOAC in patients with mitral stenosis, most of the NOAC studies excluded patients with highly thrombogenic conditions such as mitral stenosis and presence of prosthetic valve. Thus, there is little evidence of using NOAC in patients with MS to date.

Presence of mechanical prosthesis is also one of the highly thrombogenic conditions. Warfarin is routinely recommended for all patients with mechanical prosthesis to prevent thromboembolic event [61, 62]. Intensity of the treatment is different according to the site of the prosthesis. Therapeutic rage or INR 2.5–3.5 and 2.0–3.0 is recommended for mitral and aortic mechanical prosthesis, respectively [63–66]. Low-dose aspirin (75–100 mg) may have additional benefit on top of warfarin monotherapy in patients with mechanical prosthesis [67, 68].

Bioprosthesis is usually less thrombogenic than mechanical prosthesis with exception of the first 3 months after surgery. Anticoagulation may be reasonable for the first 3 months after bioprosthetic valve replacement, but supporting data is lacking [61, 62]. After 3 months from bioprosthetic valve replacement, antiplatelet therapy with aspirin is generally recommended for stroke prevention unless there is other indication for anticoagulation. One study investigating the effect of dabigatran in patients with mechanical prosthesis demonstrated unfavorable results of increased thromboembolic event and higher bleeding risk compared to warfarin [69]. There is no evidence of NOAC in patients with prosthetic valve to prevent thromboembolic events.