Embolic stroke – Embolism is especially likely in the following circumstances:

• The onset is sudden and the neurologic deficit is maximal from the beginning

• The infarct and deficit are large

• There is a known cardiac or large artery lesion present

• The infarct is or becomes hemorrhagic on CT or MRI

• There are multiple cortical or cortical/subcortical infarcts in different vascular territories

A possible cardiac source should be considered in all patients with suspected embolic stroke, particular in patients under age 45 or without clinical evidence of cardiac disease [2].

Cardiac evaluation – Transesophageal echocardiography (TEE) is the best test to exclude significant ascending aortic atheromatous disease. It is also the best way to identify clot in the left atrial appendage. These emboli are 3 mm or less in diameter and are usually beyond the resolution of the transthoracic echocardiographic ultrasound probe [2-5]. In contrast, left ventricular thrombi in patients with congestive heart failure or a previous myocardial infarction are best seen with transthoracic echocardiography (TTE) since the true left ventricular apex is not well seen on TEE [6].

Patients with suspected embolic stroke should have a TTE. A TEE is performed to examine the atria, atrial septal region, and the aorta if the TTE and preliminary cardiac and vascular imaging tests do not clarify the cause of brain ischemia [7-9]. One analysis suggested that performing TEE alone in all patients with new onset stroke was more cost-effective that performing TTE and TEE in sequence [10]. However, TEE is an uncomfortable invasive procedure that may not be tolerated by very ill patients.

There has been a preliminary report of imaging the aorta through ultrasound probes placed in the right and left supraclavicular fossae [11]. The clinical utility of this technique remains to be determined.

An electrocardiogram and Holter monitor can identify patients who have atrial fibrillation as a possible source of emboli. However, the demonstration of normal sinus rhythm does not exclude intermittent atrial fibrillation.

Vascular studies – The extracranial and intracranial arteries are also common sources of brain embolism and should be studied. 

• If the infarct and brain symptoms are within the anterior circulation (carotid artery supply), then the extracranial and intracranial carotid arteries, and the middle and anterior cerebral artery branches should be the focus of the examinations.

• When the infarct is within the posterior circulation (vertebrobasilar system), the extracranial and intracranial vertebral arteries, the basilar artery, and the posterior cerebral arteries should be the focus of the vascular investigations [12].

The anterior circulation can be studied using duplex ultrasound of the neck and transcranial Doppler (TCD) of the intracranial arteries [13-16]. B-mode images of the carotid artery also demonstrate the degree of stenosis and irregularities or ulcerations within plaques. The morphology of carotid artery plaques is well shown by duplex ultrasonography [14]. 

Alternatively, CT angiography or MR angiography of the neck and head arteries may be sufficient. Conventional angiography is performed when the screening tests do not fully define the lesion and more characterization is warranted, and when interventional treatment through an arterial catheter (angioplasty or intraarterial thrombolysis) is indicated.

Within the posterior circulation, duplex and color-flow Doppler investigation of the origins of the vertebral arteries [12,17] and ultrasound of the subclavian arteries (especially when the radial pulse or blood pressure on one side is lower than the other) can detect lesions of the proximal portion of the vertebral arteries. Atherosclerosis most often affects this proximal region. The ultrasonographer can then insonate over the rest of the vertebral artery in the neck using a continuous-wave Doppler to detect the direction of flow within the artery (craniad as would be normal, or reversed or to-and-fro flow suggesting proximal obstruction).

CTA and MRA of the neck vertebral arteries are also helpful, but these tests often do not adequately show the origins of the vertebral arteries [12,17]. The clinician must be certain that the films are adequate to see the complete intracranial vertebrobasilar system.

An important advance in the detection of brain embolism is monitoring by TCD [18,19]. Emboli that pass under ultrasound probes make a high-pitched chirp and are recorded as high-intensity transient signals (HITS). The location and pattern of these emboli can help define the presence of embolism and give clues as to the source. TCD monitoring can also help to assess the effectiveness of treatment.

Artery-to-artery versus cardiac sources of embolism – The distinction between artery-to-artery and non-artery-to-artery sources of embolism can be difficult. Suspicion of the former typically arises once vascular pathology in a large vessel has been identified (eg, with noninvasive testing). Repetitive spells within a single vascular territory are also suggestive of an artery-to-artery source, as is a normal echocardiogram. However, caution must be used in interpreting the results of a transthoracic echocardiogram. As previously mentioned, this study can exclude a cardiomyopathy and most atrial and mitral valve pathology, but may miss other potential embolic sources such as clot in the atrial appendage, a patent foramen ovale, mitral valve lesions, and aortic atherosclerosis. Transesophageal echocardiography is better for identifying these lesions.

Small vessel (lacunar) stroke – Most patients with lacunar infarcts have risk factors for penetrating artery disease (eg, hypertension, diabetes mellitus, or polycythemia). The clinical findings typically conform to one of the well recognized lacunar syndromes: pure motor hemiparesis, pure sensory stroke, dysarthria-clumsy hand, or ataxic hemiparesis.

Further testing is of low yield in patients suspected of having a lacunar infarction who have the typical risk factors, clinical neurologic findings, and characteristic brain imaging (eg, small subcortical infarct). Vascular imaging (CTA or MRA) can be performed at the same time as brain imaging (CT or MRI) to exclude occlusion of the parent feeding artery, a condition that can mimic a lacunar infarct [20,21]. It is particularly important to perform intracranial vascular imaging in blacks and persons of Asian descent, since intracranial large artery occlusive disease is common in these patients. An alternative diagnostic test to exclude intracranial occlusive disease is TCD, a technique that measures the blood flow velocities in the large intracranial arteries using an ultrasound probe placed over the orbit, temporal bone, and foramen magnum [22-24]. 

Large vessel atherothrombotic stroke – Large vessel atherothrombotic strokes are often preceded by TIAs and the onset is not abrupt. The course of neurologic symptoms and signs fluctuates or is progressive in development. Infarcts that are large and subcortical are usually caused by occlusion of intracranial arteries.

Patients with suspected large vessel atherothrombotic strokes need to have both intracranial and extracranial vascular testing. Extracranial vascular testing can be performed with MRA, CTA, or duplex carotid ultrasound. All are reliable and specific for detecting important severe occlusive lesions in the extracranial carotid arteries. Newer ultrasound devices such as color-flow Doppler imaging [25] and power Doppler [26] improve the resolution and quantification of carotid artery lesions.

One approach to patients with suspected carotid artery stenosis is to first perform carotid duplex ultrasound. Patients with stenoses less than 50 percent are followed with serial examinations, usually on an annual basis to determine if there is progression. If there is greater than 50 percent stenosis, the patient is evaluated with transcranial Doppler examination and MRA. CTA is performed in lieu of MRA if there is a contraindication to magnetic resonance imaging and in cases in which the duplex ultrasound and MRA do not agree. Conventional angiography is rarely performed; indications include patients who cannot tolerate an MRA or those in whom nonatherosclerotic disease is suspected (eg, vasculitis).

Patients with carotid artery stenosis who have had nondisabling strokes may be candidates for carotid endarterectomy. 

EVALUATION OF PATIENTS WITH INTRACEREBRAL HEMORRHAGE – CT scans usually define the size, location, and drainage pattern of intracerebral hematomas. Mass effect caused by the intracerebral mass, shift of midline structures, herniation of brain contents from one compartment to another [27,28], and the presence of hydrocephalus are also easily seen on CT. Vascular malformations and brain tumor are better visualized on MRI. The location and appearance of the lesion as well as the history of the patient (eg, race, blood pressure, presence of known bleeding disorder, use of drugs) help determine the choice of additional diagnostic studies.

No further diagnostic tests are necessary in the severely hypertensive patient with a well circumscribed and homogeneous hematoma that is located in a typical location for hypertensive ICH (eg, putamen/internal capsule, caudate nucleus, thalamus, pons, or cerebellum); the clinician can be confident that the patient has a hypertensive hemorrhage in this circumstance.Similarly, a traumatic etiology can be diagnosed with confidence in the patient who has had recent trauma and lesions in the appropriate location and with the appearance of contusion and traumatic hemorrhages (eg, anterior and/or orbital frontal lobes and temporal lobes at the surface).

Bleeding disorder – Evaluation for a bleeding disorder (platelet count, prothrombin time, and activated partial thromboplastin time) should be performed in every patient with an intracranial hemorrhage, especially if the cause is not immediately clear. A bleeding tendency can cause or contribute to bleeding initiated by other etiologies.

Iatrogenic prescription of anticoagulants is the most common bleeding disorder contributing to brain hemorrhages. In one series of 24 such patients, 18 were hypertensive; only one had simultaneous bleeding in other organs [29]. These bleeds are most often lobar or cerebellar. Anticoagulant hemorrhages often develop gradually and may become progressively larger over hours or even a few days [29,30].

Lobar or atypical hemorrhage – Amyloid angiopathy, bleeding into a tumor, and vascular malformations are likely etiologies of hemorrhages that are lobar or atypical in appearance.

Cerebral amyloid angiography – Hemorrhages related to amyloid angiopathy are usually lobar, but are occasionally cerebellar [31,32]. They predominantly involve the posterior portions of the brain, including the parietal and occipital lobes. The hemorrhages are usually multiple; gradient-echo MRI may show the presence of old small hemorrhages. Patients with amyloid angiopathy are typically over the age of 65. 

Vascular malformation or brain tumor – Other bleeding lesions should be excluded in patients under the age of 60 if the blood pressure is not sufficiently elevated to make a firm diagnosis of hypertensive lobar hemorrhage. A repeat MRI after the blood has been reabsorbed (four to eight weeks) may show residual vascular malformations or a brain tumor. Vascular imaging using a helical (spiral) CT scan and CTA or MRA of the intracranial circulation are useful screening tests for vascular malformations and aneurysms [33,34]. Contrast angiography by arterial catheterization may be necessary in patients with a CTA or MRA suggestive of vascular malformation.

OTHER CONSIDERATIONS – A cardiac evaluation is important in virtually all patients with brain ischemia [35,36]. Not only are cardiac and aortic embolism common, but many patients with cerebrovascular occlusive disease have concurrent coronary heart disease that can lead to significant morbidity and mortality. A thorough history focusing on the presence of cardiac ischemia and arrhythmias, a careful cardiac examination, and an electrocardiogram (ECG) are important in every patient. Many will also need echocardiography (see above).

A number of blood tests are indicated in patients with brain ischemia, including [37]:

• Complete blood count, including hemoglobin, hematocrit, white blood cell count, and platelet count

• Prothrombin time and partial thromboplastin time

• Blood lipids, including total, HDL, and LDL cholesterol, and triglycerides.

Plasma fibrinogen levels are important predictors of coronary heart disease and stroke [38-40]. However, the routine use of fibrinogen as a cardiovascular risk marker is limited by measurement variability. If fibrinogen is used clinically, three measurements should be obtained at each venipuncture and a minimum of two sets of fibrinogen measurements are needed. An elevated fibrinogen concentration may make a difference in the selection of cholesterol lowering medications in patients with hyperlipidemia. 

Other studies may include hemoglobin electrophoresis in patients suspected of having a hemoglobinopathy, and an erythrocyte sedimentation rate, tests for Lyme disease, syphilis, and HIV infection in selected patients. Hyperviscosity and hypercalcemia [43] can also contribute to or cause brain ischemia; serum protein levels (albumin, globulins), serum electrophoresis, and serum calcium, phosphate, and alkaline phosphatase are important measurements in some patients.

Measurement of platelet adhesion molecules as a marker of platelet activation may prove to be useful for monitoring antiplatelet therapy and distinguishing between patients with thrombotic and cardioembolic strokes. Platelet activation has an important role in the pathogenesis of coronary heart disease. A similar situation may occur in individuals with acute cerebral ischemia due to large vessel atherosclerotic disease. Support for this hypothesis was derived from one study that utilized flow cytometric techniques to detect platelet adhesion molecules in 72 subjects with acute cerebral ischemia (either stroke or TIA) and controls [44]. Platelets from the patients who had a thrombotic event expressed significantly more activation markers than those in controls; there was no similar difference between patients with cardioembolic events and controls. The clinical application of this information is not clear at present; further study is necessary.