Clinical manifestations

Presentation and course

Posterior circulation infarcts comprise 20% to 30% of all ischemic strokes (79; 106; 109; 32). Vertebrobasilar ischemic stroke is not a homogenous entity; a complexity of neurologic symptoms, distinct stroke patterns, clinical courses, and outcomes may be observed (26).

The duration of transient ischemic attacks varies considerably, although most usually resolve one hour after onset (87; 83; 62). Symptoms vary from mild to extreme, and the frequency may vary from a solitary attack to many per day when evaluated over a period of years (26; 40).

Vertebrobasilar transient ischemic attacks often present with unilateral motor or sensory symptoms involving the face or limbs as well as unilateral or bilateral visual field defects. Occasionally, motor or sensory symptoms may be bilateral or change from one side to the other between attacks (61). Vertebrobasilar transient ischemic attacks may be accompanied by vertigo, diplopia, dysarthria, dysphagia, or loss of balance, often in association with any combination of motor or sensory deficits in the face and extremities (79). However, there are sudden onset, nonprogressive, nonfocal events (nonconsensus or atypical transient ischemic attacks) that occur in isolation, sometimes days or weeks before ischemic stroke, and that do not meet the traditional criteria, such as vertigo, unsteadiness, ataxia, partial sensory deficits, dysarthria, and visual deficits, including diplopia and bilateral decreased vision. Nonconsensus transient ischemic attacks accounted for nearly 20% of 2878 patients with minor ischemic stroke and transient ischemic attacks in the Oxford Vascular Study, a longitudinal, population-based, prospective cohort study between 2002 and 2018 (97). Isolated vertigo and ataxia, bilateral decreased vision, and isolated diplopia were reported in 37%, 17%, and 8% of patients, respectively, with nonconsensus transient ischemic attacks suggesting a posterior circulation involvement. Compared to patients with classic transient ischemic attacks, those with atypical transient ischemic attacks were younger (mean age 68 years) and had significantly fewer vascular risk factors, a lower history of previous stroke, and lower ABCD2 scores (mean 2.8 points). Moreover, patients with nonconsensus transient ischemic attacks had double the posterior circulation stenosis rate of those with classic transient ischemic attacks (18% vs. 9%). A delay in seeking medical attention occurs in the nonconsensus transient ischemic attack scenario, and the initial assessment tends not to consider a vascular etiology of such atypical symptoms, resulting in a substantially lower prescription of secondary stroke prevention measures than in classic transient ischemic attack cases (79; 62; 97).

Finally, isolated loss of consciousness, incontinence, confusion, and amnesia are not usually considered vertebrobasilar transient ischemic attacks, and in these cases, other causes should be ruled out (03). Notably, in 7 of 13 (54%) patients with ischemic amnesia, amnesia symptoms lasted less than 24 hours. A third of these patients had precipitating stroke factors, but none had emotional or physical precipitating factors. Involvement of the Papez circuit or associated structures was found on diffusion-weighted MR images in 85% of patients (72). The posterior circulation was involved in 12 patients, four of whom also had simultaneous anterior strokes. Half of all strokes were of cardioembolic etiology.

Prognosis and complications

Transient ischemic attack should be considered a medical emergency. Formerly, the risk of stroke in the short aftermath of a transient ischemic attack was thought to be low; however, data have found a high risk of stroke soon after transient ischemic attack. In a prospective population-based study, a stroke risk of 5% was found within 24 hours after a first transient ischemic attack. Half of all recurrent strokes occurring within seven days after a transient ischemic attack happened in the first 24 hours, and most were disabling or fatal (12). The current prognosis for patients with transient ischemic attacks is reportedly more favorable than previously thought. A meta-analysis of 68,563 patients with a transient ischemic attack diagnosis included in 40 studies published from 2008 to 2015 found cumulative stroke risk following a transient ischemic attack to be 1.2%, 3.4%, 5.0%, and 7.4% at 2, 7, 30, and 90 days, respectively (73). There were considerable drawbacks in studies included in this meta-analysis, and it is uncertain whether its findings reflect a better approach to transient ischemic attack patient care.

A current estimation of early stroke risk post-transient ischemic attack was performed in a meta-analysis of 206,455 patients (mean age 69 years, 58% women) from 25 countries, most of them industrialized nations (88). All of these patients were included in 68 studies between 1981 and 2018; in 31 of the studies, transient ischemic attack patients were recruited after 2007, and eight of these latter trials used a tissue-based transient ischemic attack definition. The pooled stroke risk following a transient ischemic attack was 2.4%, 3.8%, 4.1%, and 4.7% within 2, 7, 30, and 90 days, respectively (88). Studies over the last two decades showed a lower early stroke risk than those prior to 1999, which might reflect either proper exclusion of minor stroke patients or an improvement in secondary prevention.

Traditionally, patients with transient ischemic attacks in the posterior circulation have been considered to have a favorable prognosis. Nevertheless, a systematic review of 37 published cohort studies found a higher 7-day stroke risk in patients with transient ischemic attacks or minor stroke involving the posterior circulation than in those patients in whom the carotid circulation was involved (29).

In a pooled individual patient analysis of two prospective stroke studies in 359 patients with vertebrobasilar transient ischemic attack or cerebral infarct, patients with intracranial vertebral artery stenosis had a higher risk of stroke recurrence in the posterior circulation at 90 days than those with extracranial vertebral stenosis or no stenosis (36). A cut-off of more than 50% reduction in lumen diameter in the extracranial or intracranial vertebral artery was determined by CTA, contrast-enhanced MRA, or both. The risk of stroke (transient ischemic attack or cerebral infarct) recurrence was four times higher in patients with intracranial vertebral stenosis than in those with no vertebral stenosis. Patients with extracranial vertebral stenosis had a 2-fold higher risk of new clinical stroke in the vertebrobasilar territory than patients with no vertebral artery stenosis. Vertebral stenosis was independent of traditional risk factors as a predictor of recurrence.

The conventional wisdom is that patients with nonconsensus transient ischemic attack symptoms have a lower stroke risk compared to patients with typical transient ischemic attacks. A prospective study of patients with minor ischemic stroke, classic transient ischemic attacks, or nonconsensus transient ischemic attacks showed that patients with the latter have a prevalence of cardiovascular pathological findings (atrial fibrillation, patent foramen ovale, and any arterial stenoses greater than 50% at baseline) similar to that of patients with classic transient ischemic attacks, although stenoses in the posterior circulation were more frequent in patients with nonconsensus transient ischemic attacks. No differences were found regarding the frequency of acute ischemic lesions on diffusion-weighted imaging between patients with classic transient ischemic attacks involving the posterior circulation and patients with nonconsensus transient ischemic attacks (97). In the prospective study in question, the median follow-up was 5.2 years, and brain infarction in the short- and long-term was found to be as frequent in patients with nonconsensus transient ischemic attacks as in those with classic transient ischemic attacks. A recurrent ischemic stroke after the index event and before seeking medical attention was significantly more common in nonconsensus patients (8%) than in classic patients (5%). The 90-day stroke risk was 10.6% and 11.6% for atypical and classic patients, respectively. The 10-year risk of all major vascular events was 27.1% for nonconsensus patients and 30.9% for classic patients. Considering these findings demonstrating the prevalence of atypical transient ischemic attack symptoms before vertebrobasilar ischemic stroke and the respective high early and long-term stroke risk, the authors called for a revision to the list of transient ischemic attack–consistent symptoms. The designation of nonconsensus transient ischemic attacks as definite cerebrovascular events will increase overall transient ischemic attack diagnoses by approximately 50% (97).

Front-line emergency department physicians or neurologists (or both) evaluating patients with transient ischemic attacks either in an emergency room or stroke prevention clinic setting need reliable and simple methods to stratify individual stroke-risk profiles to screen patients for urgent assessment and make appropriate clinical decisions. A 6-point ABCD score has been proposed and validated to identify patients with an early high risk of stroke after a transient ischemic attack. The score is based only on clinical parameters: age, blood pressure, clinical features (weakness and speech impairment), and spell duration. Patients with a score of 6 showed a 30% risk of stroke within seven days after the index transient ischemic attack (81). This score has been modified by the addition of a point for diabetes--ABCD2 (49). Refinements of prediction rules based on the ABCD2 score have been proposed. A triage algorithm (the ASPIRE approach) was assessed, albeit retrospectively, in approximately 570 patients with a transient ischemic attack diagnosis at a tertiary care referral center to identify patients at high risk of further stroke (20). In this approach, patients determined to be at high risk (6.3% 90-day risk of stroke) were those having any one of the following components: ABCD2 score 4 or more points, weakness or language disturbance lasting longer than 5 minutes, or atrial fibrillation. Moreover, the addition of a biological marker – the reactive C protein levels – to ABCD2 score may help to predict vascular recurrence. In the first hours after a transient ischemic attack, patients with reactive C protein levels above 3 mg/l (excepting patients with concurrent infection or inflammatory conditions) and ABCD2 score 4 or more points had a higher risk of further ischemic stroke than those who had been stratified based exclusively on clinical findings (18). Several clinical-plus scores have been developed specifically for use in specialized stroke centers to enhance the discriminative ability of clinical scores to stratify high-stroke-risk patients. Along with clinical findings, the clinical-plus scores include neuroimaging findings (CT, diffusion-weighted MRI) and etiological mechanisms (extracranial or intracranial carotid stenosis, cardiac embolism, or undetermined etiology). Thus, it has been suggested that radiological and, particularly, diffusion-weighted MRI findings improve the ability of the ABCD2 clinical score to discriminate low- or high-risk of further stroke (19). Accordingly, the ABCD3-I score (a refined clinical and imaging-based prediction score having 13 points instead of the classical 7-point ABCD2) assigns two points each for a “dual” transient ischemic attack (a recurrent transient ischemic attack within 7 days after an index event), positive diffusion-weighted imaging, and stenosis of over 50% on carotid imaging (71). This score substantially improved predictive ability and risk classification compared to the ABCD2 score. A refinement of the ABCD3-I score has been proposed involving the replacement of the dual transient ischemic attack item by dual DWI (diffusion-weighted imaging lesions of different age or in both carotid territories and the vertebrobasilar system). The dual diffusion-weighted imaging component has proven to show a better and independent predictive ability for further stroke at 90 days than dual transient ischemic attack in transient ischemic attack patients (22). Both ADBCD3-I scores, the first and the most recent, included a stenosis of over 50% exclusively in the carotid artery, with no consideration of any significant stenosis in the vertebrobasilar system. Nevertheless, the most recent ABCD3-I score included the presence of lesions in the posterior circulation determined by MRI-DWI. To date, it seems that the different transient ischemic attack scores designed to stratify the risk of subsequent early stroke are most applicable to patients with anterior circulation symptoms. In fact, in 369 patients with a transient ischemic attack diagnosis, the ABCD2 score was less valuable in predicting a subsequent stroke 7 days later in patients with a posterior circulation than when the transient ischemic attack occurred in the anterior circulation (100). In a pooled analysis of 2176 patients in 16 heterogeneous cohort studies who had recently had a transient ischemic attack confirmed by a stroke specialist, the ABCD3-I score had a better predictive ability in acute hospital settings to discriminate patients with high early stroke risk after a transient ischemic attack as compared with the ABCD2 score (53). The analysis unfortunately failed to differentiate between anterior and posterior circulation transient ischemic attacks.

Moreover, in a retrospective series of 129 transient ischemic attack patients, proximal intracranial stenosis (greather than 50%) or occlusion on MRA within 24 hours of the transient ischemic attack was predictive of a high risk for early stroke recurrence (transient ischemic attack or stroke), regardless of the ABCD2 score (90). The rate of early stroke recurrence (all strokes occurred within 48 hours after the index event and in the stenotic/occluded artery territory) was 2.5 times higher for patients with significant steno-occlusive lesions on proximal intracranial vessels (including basilar and vertebral arteries) than for patients with an ABCD2 score of greater than 4 (10% rate of early stroke recurrence). The correlation was weaker in patients with steno-occlusive lesions located at any site (proximal or distal). There were no recurrences in patients with an ABCD2 score of less than 4.

In a small multiethnic cohort of 148 hospitalized patients with transient ischemic attack as the qualifying event, stroke risk in the first 7 and 90 days was higher in patients with the highest ABCD2 scores (6 and 7 points) (96). When the patients’ analyses were dichotomized between low risk (0 to 3 ABCD2 points) and high risk (4 to 7 ABCD2 points), the latter group had a 6-fold higher stroke risk during the next 90 days. This finding supports expediting diagnostic evaluation and treatment in patients with an ABCD2 score of 3 and above. Nevertheless, patients with an ABCD2 score of less than 4 may also have a high short-term stroke risk, as has been demonstrated in a prospective cohort of 1679 patients with definite or possible transient ischemic attack. The 90-day stroke rate was 3.9% in 180 patients with an ABCD2 score of less than 4 when at least one criterion for emergency treatment was present (major cardiac source of embolism and symptomatic internal carotid stenosis or intracranial stenosis greater than 50%), whereas the 90-day stroke rate was 3.4% in 701 patients with an ABCD2 score of greater than 4 and no criteria for emergency treatment (02).

ABCD3-I score performance was evaluated in the emergency setting in patients with a suspected transient ischemic attack. In the first study phase, a pre-implementation cohort comprised 143 patients who underwent neuroimaging studies at the discretion of the treating physician; in the second phase, a postimplementation cohort of 118 patients was assessed through standard protocol using the ABCD3-I algorithm (21). Patients in the pre-implementation cohort were mean aged 68. Their mean ABCD2 score was 4.5 points; 92% were admitted to the hospital and the remainder sent home. Patients in the postimplementation cohort were mean aged 64. Their mean ABCD2 and ABCD3-I scores were 4.1 and 4.5 points, respectively; 76% were discharged to home (ABCD3-I less than 7 points and negative diffusion-weighted magnetic resonance imaging), and the remainder admitted to the hospital if they had ABCD3-I of 8 points or greater or ABCD3-I of 7 points or less with abnormal diffusion-weighted imaging. Clinical outcomes at 90 days after index transient ischemic attack showed no significant differences in ischemic stroke rate between the two cohorts (21). Thus, in emergency settings, the ABCD3-I algorithm may reduce unnecessary hospital admissions and health care costs for transient ischemic attack patients without adversely affecting clinical outcome prediction.

In a prospective cohort of 723 Chinese patients (mean age 62 years) with noncardioembolic infarcts in the posterior circulation, the rate of stroke recurrence within one year after index event was reported for both infarcts and transient ischemic attacks (107). Forty percent of infarcts involved the middle posterior circulation territory, which comprises mainly the pons. The distal posterior circulation territory accounted for another 40% of infarcts, which are supplied by the rostral basilar artery, the superior cerebellar artery, and the posterior cerebral artery, including their penetrating artery branches. In this cohort, 20 patients (2.8%) had previous repeated transient ischemic attacks three months before the qualifying posterior circulation infarct, and 35 patients (4.8%) had a transient ischemic attack in the vertebrobasilar circulation within one year after the index event. The predictors of stroke recurrence were dysphagia at admission, repeated transient ischemic attacks before infarct onset, multisector infarcts, and stenosis higher than 70% in the symptomatic artery (107).

The 5-year risk of stroke, acute coronary syndrome, or death from cardiovascular causes – whichever occurred first – was assessed in an international multicenter trial of 3417 patients (mean age 66 years) with transient ischemic attack or minor stroke within seven days prior to enrollment (04). Two thirds of all patients included in this trial had a moderate to high risk of stroke (4 points or more on the ABCD2 score at baseline). At the end of follow-up, the estimated cumulative rate of cardiovascular events was 12.9%; half of these events occurred between the second and fifth year. At 5 years, the risk of stroke and transient ischemic attack was 9.5% and 8.3%, respectively. Among 345 patients with stroke recurrence, 57% of the events occurred within the first year after the index event and 13% of patients had a fatal stroke within the 5-year follow-up period (04).

Clinical vignette

Seven days before admission, a right-handed 67-year-old man with a history of arterial hypertension and hypercholesterolemia experienced a sudden onset of dysarthria, vertigo, and weakness in his left arm and leg lasting 15 minutes, with subsequent restoration of normal muscular strength. At the time, he was taking 12.5 mg of chlorthalidone every other day. He had experienced no previous cerebral infarct or transient ischemic attacks.

Five days later, he had a second episode of sudden onset of dysarthria with vertigo, blurry vision, and weakness on his left side. These symptoms resolved completely within 20 minutes during the ambulance ride to the hospital.

On admission, his supine blood pressure was 160/95 mm Hg with a regular heart rate of 80 beats per minute. His neurologic examination was normal, except for diminished hearing in both ears. A laboratory evaluation consisting of a complete blood count, erythrocyte sedimentation rate, prothrombin and partial thromboplastin times, VDRL test, and routine blood chemistries was also normal. A standard 12-lead electrocardiogram showed a normal sinus rhythm, but a lipid profile revealed hypercholesterolemia. A transthoracic echocardiogram disclosed left ventricular hypertrophy and no cardiac sources of embolism. Brain magnetic resonance imaging, done the following day with diffusion-weighted imaging included, was found to be normal. A magnetic resonance angiography showed a mild basilar stenosis.

The patient was put on aspirin, enalapril, and atorvastatin and discharged. Five months later, he remained symptom-free, with no further ischemic episodes.

Biological basis

• Transient ischemic attacks tend to be more common in patients with either vertebral (extracranial and intracranial) or basilar artery disease.

Etiology and pathogenesis

Transient ischemic attacks are commonly found in different vascular conditions associated with posterior circulation ischemia and tend to be more common in patients with either vertebral (extracranial and intracranial) or basilar artery disease than in those with intrinsic posterior cerebral artery disease. Vertebral or basilar stenosis greater than 50% was found in 26% of a cohort of 141 patients with transient ischemic attacks or minor stroke in the posterior circulation; 60% of these stenoses were located in the extracranial artery, and a third were in the intracranial segment of the vertebral artery (70). Moreover, symptomatic vertebrobasilar stenosis was associated with multiple transient ischemic attacks shortly before medical attention was sought, as well as with a high risk of early recurrence--45% had a recurrent transient ischemic attack or infarct within 90 days after the index event (70).

Transient ischemic attacks were the index event in 70% of 39 patients with symptomatic ostial stenosis (greater than 70%) of the vertebral artery with no other identifiable cause, who were followed to assess long-term outcomes in severe occlusive vertebral disease (51). A transient ischemic attack may occur in as many as one third to two thirds of patients with moderate to severe basilar artery stenosis or occlusion, whereas posterior cerebral artery occlusions are less prone to transient ischemic attacks (26). Isolated symptomatic basilar stenosis (at least 50% of basilar narrowing assessed by noninvasive or invasive angiography methods) was found in 91 (1.43%) of 6369 patients with acute ischemic stroke (85). Midbasilar stenosis predominated, mean basilar stenosis was 81%, and the mean length of stenosis was 4.7 mm. Among patients with basilar stenosis, 20% had a transient ischemic attack as the index event, and the rest had a cerebral infarct (85).

Dolichoectasia—a little understood arteriopathy characterized by abnormal dilation, elongation, and tortuosity of at least one intracranial vessel—usually affects older individuals with cardiovascular risk factors. Intracranial dolichoectasia prevalence in stroke populations is in the 3% to 17% range; the basilar artery is predominantly affected, and transient ischemic attack or cerebral infarct may be the initial clinical presentations in about 40% of patients with vertebrobasilar dolichoectasia (74).

Spontaneous or traumatic cervical artery dissection is a common cause of ischemic stroke in young adults and the middle-aged, though it may occur at any stage of life. Vertebral artery dissections are less common than those in the carotid artery, but a quarter of all cerebral infarcts in the posterior circulations may be due to dissection of vertebral arteries. Intracranial dissections of the vertebral artery are more common than extracranial dissections and often extend to the basilar artery (13). Vertigo, occipital headache, and ipsilateral neck pain are the most frequent clinical manifestations of vertebral artery dissections; transient ischemic attacks occurred in 25% of 153 patients with vertebral artery dissections, most often when extracranial dissection extended to the intracranial vertebral segment (27).

The underlying pathophysiological mechanisms behind ischemia in transient ischemic attacks do not differ from those of cerebral infarcts. Therefore, the primary process may lie in the large or small arteries of the vertebrobasilar territory, in the heart, or in a blood coagulation disorder (34). Atherosclerosis, the main pathology behind large artery disease, predominates in the intracranial arteries of the vertebrobasilar circulation. Steno-occlusive atheromatous disease in the posterior circulation is commonly located at the origin of the vertebral artery and in the distal intracranial vertebral arteries, the proximal and middle basilar artery, and the proximal posterior cerebral artery (69). A thrombus may superimpose over the atheromatous plaque, mainly when ulcerated, and this may cause in situ thrombotic occlusion, branch occlusion, artery-to-artery occlusion, or hemodynamic compromise (if there is severe coexisting stenosis in the appropriate artery). In situ thrombotic occlusions often occur in the posterior cerebral artery, anterior inferior cerebellar artery, or posterior inferior cerebral artery. An atheromatous plaque in an intracranial artery can occlude one or several perforators (branch artery disease) and cause ischemia (69). This mechanism may be associated with infarcts in the penetrating arteries of the thalamus, pons, and medulla oblongata. Moreover, an embolus from a platelet-fibrin thrombus attached to the atheroma in a proximal artery may easily dislodge and migrate to mechanically block distal intracranial arteries (34). Common recipient sites for posterior circulation embolism are the intracranial vertebral artery, posterior inferior cerebellar artery, distal basilar artery, and posterior cerebral artery. Small artery disease due to lipohyalinosis (intrinsic disease) of a penetrating artery preferentially affects arteries between 40 to 400 microns in diameter, mainly of the basilar or posterior cerebral artery (34). About 20% to 25% of symptomatic cardiac emboli to the brain lodge in the posterior circulation; an embolus from a cardiac source leading to a vertebrobasilar transient ischemic attack may be due to nonvalvular atrial fibrillation, ventricular dyskinesia, valvular heart disease, patent foramen ovale, endocarditis, and auricular myxoma. The most common sites for cardiac embolism are the extracranial vertebral posterior inferior cerebellar artery system, followed by the top of the basilar artery, the superior cerebellar artery, and the posterior cerebral artery. Several mechanisms may sometimes be present, making it difficult to define the putative mechanism responsible for the transient ischemic attack.

The leading pathophysiological mechanism in 240 atherosclerosis patients with posterior circulation ischemic stroke was artery-to-artery embolism (43%), followed by local branch occlusion (34%), in situ thrombotic occlusion (12%), and mixed mechanism (10%) (55).

Less common causes of transient ischemic attack include nonatherosclerotic vasculopathies, such as vertebral dissection (often in young adults), fibromuscular dysplasia, dilatative arteriopathy (dolichoectasia), infectious or immune CNS vasculitis, and subclavian steal syndrome. Blood and coagulation disorders are an uncommon cause of transient ischemic attack, though they should be considered in individuals under the age of 50 without cardiovascular risk factors and patients with a history of clotting dysfunction.

The leading etiology in 2873 Asian patients (mean age of 64 years) with a transient ischemic attack (time-based definition) or minor stroke (less than 4 points in the NIHSS score at baseline) and positive diffusion-weighted image within seven days of stroke onset was large artery atherosclerotic disease (42.5% according to TOAST criteria), followed by small vessel disease (23%) and embolism from a cardiac source (12%), whereas etiology could not be determined in 18% of the patients (56). Around one-third of patients presented with relevant arterial disease (defined as stenosis greater than 50% of vessel diameter or occlusion in the appropriate arterial territory assessed by magnetic resonance angiography), computed tomography angiography, or both.

Subclavian steal syndrome. Subclavian steal syndrome is due to occlusion or high-grade stenosis of the innominate or the subclavian artery in the proximal segment before the vertebral artery origin, resulting in alteration of vertebral artery hemodynamics (80). In this setting, the low pressure in the subclavian system distal to the lesion eventually results in reversal of flow in the ipsilateral vertebral artery, which serves as a source of collateral flow to the distal subclavian artery.

A 5.4% prevalence of subclavian steal syndrome was found in nearly 8000 patients who had a Doppler ultrasonography of extracranial neck vessels (60). The subclavian steno-occlusion is usually unilateral and very rarely bilateral (103). The most common causes are atherosclerosis of the subclavian artery in those over 50 years of age and Takayasu arteritis in young people, particularly of Asian descent (25; 103). Due to the more acute angle of takeoff and consequent turbulent flow, which increases the risk of atherosclerosis at the subclavian-aortic junction, the left subclavian artery is more frequently affected than the right (25; 60; 50). Other conditions, though uncommon, that can also cause subclavian steal syndrome are external compression of the subclavian artery at the thoracic outlet, giant cell arteritis, dissection, stenosis after surgical repair of aortic coarctation or tetralogy of Fallot, and traumatic or congenital anomalies of the innominate or subclavian arteries (67).

The hemodynamic alteration in the subclavian artery may remain asymptomatic (“subclavian steal phenomenon”), or it might become symptomatic if hemodynamic compensation fails, which may occur when the subclavian stenosis reaches a critical level and there is either a significant concomitant carotid disease, a posterior communicating artery stenosis, or an inadequate intracranial collateral flow (80). A steal phenomenon is often elicited or aggravated by vigorous use of the ipsilateral upper limb to subclavian stenosis, thereby increasing the flow requirements in the ipsilateral upper arm, which leads to symptoms due to insufficient perfusion in the brain, upper arm, and occasionally in the heart. Symptoms occur more frequently when subclavian stenosis is bilateral rather than unilateral. Although it is uncommon, a bilateral symptomatic subclavian steal syndrome may occur without subclavian/innominate artery stenosis. In a 70-year-old patient undergoing dialysis for chronic renal failure, sonography demonstrated a global hemodynamic perturbation in the anterior and posterior circulation due to a high-flow arteriovenous graft that had been introduced after a preexisting arteriovenous fistula in the contralateral arm became nonfunctional (52). The outstanding clinical manifestation was recurrent syncope-like episodes of loss of consciousness while standing or sitting.

Neurologic symptoms, including paroxysmal vertigo, cortical visual disturbances (diplopia, blurred vision, or hemianopia), dysarthria, ataxia, drop attacks, and syncope, have been reported in approximately one third of patients with severe steno-occlusion of the subclavian artery (50). Sudden and excruciating headache in dialysis patients with arteriovenous or cluster-like headache is an even more unusual manifestation of subclavian steal syndrome (17).

Subclavian steal syndrome has traditionally been considered an uncommon cause of posterior circulation ischemia. This presumption has been assessed in 54 patients with subclavian steal syndrome determined by transcranial Doppler ultrasound. Forty-three patients had a stenosis over 70% of vessel diameter or occlusion of the subclavian artery as measured by CT, MR, or digital subtraction angiography, and the remainder had moderate stenosis ranging from 50% to 70% of vessel diameter (50). If a concomitant severe stenosis (over 70%) or occlusion of the vertebrobasilar artery was determined by angiography, the patients were excluded to avoid the interference of other potential pathophysiological stroke mechanisms. Subclavian steal syndrome was an incidental finding in 66% of patients; the remaining one third (18 patients) had symptoms in the vertebrobasilar circulation; 13 had a transient ischemic attack; four had an acute infarct (tiny infarcts in the pons and occipital lobe, two large cerebellar infarcts); and one patient with Takayasu arteritis developed a hypertensive encephalopathy and concomitant vasogenic edema (50). Interestingly, the severity of a subclavian steal phenomenon graded by transcranial Doppler ultrasound did not correlate with the occurrence of symptoms; instead, simultaneous atherosclerosis in other cerebral arteries and unstable cerebral hemodynamics seem to have played a determining role in the occurrence of posterior circulation ischemia in patients with subclavian steal syndrome. Hence, 87% of patients with transient ischemic attacks also had large artery atherosclerosis (mainly extracranial internal carotid artery stenosis), and 70% had unstable cerebral hemodynamics on transcranial Doppler ultrasound in the anterior circulation. Moreover, symptomatic patients had a rate of old silent infarcts in the posterior circulation 6-fold higher (50% vs. 8%) than that of asymptomatic patients (50).

In a retrospective study, the stroke rate was estimated in 102 (62%) of 165 patients with subclavian steal syndrome who demonstrated partial (72%) or complete (28%) subclavian artery occlusion on MRA, DSA, or ultrasound (94). Approximately one third of these patients were treated invasively (mostly with stents), and stroke—involving the anterior and posterior circulation in 42% and 14% of cases, respectively, and of unknown territory in 40%—occurred in 42% of the patients during a follow-up of 28 months. A significant stroke predictor was the presence of symptoms at the presentation of subclavian steal syndrome. Stroke was associated with the severity of carotid or vertebral stenosis and subclavian artery occlusion (94).

Ipsilateral arm claudication (exercise-induced pain or fatigue) may be the most common complaint when the blood flow in the upper arm is insufficient; occasionally, paresthesias may be noted at rest or with exertion. Myocardial ischemia may result from proximal subclavian stenosis after coronary artery bypass surgery (the so-called “coronary-subclavian steal syndrome”). This syndrome usually manifests as refractory unstable angina, and less frequently as myocardial infarct or ventricular arrhythmias. In this situation, the internal mammary artery is often anastomosed with the left coronary artery to perfuse the heart; hence, the blood flow in the internal mammary artery depends on sufficient arterial pressure in the subclavian artery. Consequently, a proximal subclavian stenosis may reverse the flow through the patent graft, which diverts flow from the left anterior descending coronary artery to the subclavian circulation, resulting in myocardial ischemia (80).

On physical examination, subclavian steal may manifest with an audible bruit at the base of the neck or over the supraclavicular fossa on the affected side, reduced blood pressure in the ipsilateral upper extremity (a difference in systolic blood pressure of at least 20 mm Hg between both arms), along with weakness, coolness, or absence of the radial or ulnar pulses in the affected arm.

Color-coded Doppler or transcranial Doppler ultrasound is a useful bedside screening test when a steal phenomenon is suspected (80). Both noninvasive imaging modalities are readily accessible, inexpensive, and can be repeated as needed; however, these are operator-dependent techniques. They are useful in assessing the severity of vertebral steal hemodynamics, as well as disclosing other lesions in the neck vessels. Moreover, transcranial Doppler ultrasonography can also characterize the flow patterns in the distal vertebral artery and basilar artery; thus, it is probably more useful in patients with neurologic symptoms. The hyperemia-ischemia cuff test may be used during ultrasound examination to reveal any occult steal, especially in the early stages of subclavian stenosis (50). Increasing the collateral flow in the upper arm is usually induced by inflating the arm blood pressure cuff to at least 20 mm Hg above the systolic blood pressure for 3 to 5 minutes. An inversion of the flow direction in the homolateral vertebral artery occurs, which induces reactive hyperemia and, consequently, the steal phenomenon on the release of compression.

Diagnosis must be confirmed by CT or MR angiography, because complete retrograde vertebral flow on ultrasound, especially when it is intermittent, is not always associated with a subclavian steal phenomenon, and it can also be due to proximal steno-occlusion in the vertebral artery (14). Both angiograms allow quantifying the degree of subclavian stenosis, defining the adjacent vascular anatomy, and planning interventions. However, false negatives have been reported with CT angiography because of inadequate technique and postprocessing methods. Additionally, time-of-flight MRA is a flow-sensitive technique that, however, cannot always reliably differentiate reversed flow from lost flow. Currently, the use of conventional subtraction angiography, despite its high sensitivity and specificity, has substantially decreased due to the use of less invasive techniques; however, it may play a noteworthy role in endovascular therapy.

Management of incidental subclavian steal syndrome should focus on optimal atherosclerotic risk factor control. Serious symptomatic subclavian lesions justify endovascular or surgical treatment, which should be directed to restore antegrade vertebral flow in the affected vertebral artery and improve intracranial blood flow. The first line of treatment is the endovascular approach (percutaneous transluminal angioplasty or stenting), but open surgical reconstruction is a valid alternative when the lesion is not amenable to endovascular therapy. To date, no prospective randomized studies have been performed to compare the effectiveness and safety of percutaneous transluminal angioplasty and stent placement in subclavian artery stenosis (43). In a widely cited study of a cohort of 110 symptomatic patients (55% with neurologic symptoms), the technical success rate of angioplasty was 100% in patients with proximal subclavian stenosis but dropped to 65% in patients with subclavian occlusion. The patency rate at 5 years was 89%; nearly 60% of patients had percutaneous angioplasty and stenting of the subclavian artery, and the remainder had proximal subclavian angioplasty alone. Restenosis occurred in 7% of patients over 26 months of follow-up. The combined rate of stroke and death was 3.6%; four patients had ischemic events at the end of the procedure or after the procedure (25). Higher technical success rates of over 95%, as well as patency rates of 98% and 82% at 1 and 5 years, respectively, have been reported in patients who underwent angioplasty and stenting (101). In the last few months, 21 (91%) of 23 patients with symptomatic subclavian artery stenosis had a successful outcome after undergoing endovascular treatment (balloon angioplasty and stent implantation) through the antegrade or retrograde approach, or both. The rate of recanalization for the anterograde and retrograde approaches was 68% and 75%, respectively (75). Ninety-five percent of patients had clinical symptom remission; one patient developed an arterial dissection (proximal segment of the subclavian artery) during the procedure with no permanent neurologic deficit; and in-stent restenosis occurred in three patients after a median follow-up of three years.

Moreover, patients with subclavian steal syndrome and concomitant hemodynamic significant carotid stenosis should undergo carotid treatment prior to any treatment for subclavian stenosis. Surgical options include carotid-subclavian bypass, axillo-axillary bypass, and carotid-subclavian transposition.

Epidemiology

Epidemiological data related to vertebrobasilar transient ischemic attacks are difficult to interpret, as these transient ischemic attacks are not always considered separately in the literature from vertebrobasilar stroke or from strokes or transient ischemic attacks of the carotid territory. Vertebrobasilar transient ischemic attacks account for 20% to 25% of all transient ischemic attacks (100; 23).

In a cohort of posterior circulation stroke published in the last few years, 167 (41%) of 407 patients had a transient ischemic attack, 65 (16%) only experienced a transient ischemic attack, whereas 98 (24%) of the patients developed a subsequent posterior circulation infarct after their transient ischemic attack (86). In a prospective population-based study of stroke incidence assessing the frequency of transient ischemic attacks and transient neurologic symptoms (presumably caused by cerebral ischemia) preceding vertebrobasilar stroke, 5 of 59 patients (8%) had a clinically definite transient ischemic attack according to NINDS (National Institute of Neurological Disorders and Stroke) criteria, whereas the remaining 54 patients showed transient isolated brainstem symptoms (most often isolated vertigo, although vertigo with nonfocal symptoms, isolated diplopia, transient generalized weakness, and binocular visual disturbance were also observed) before posterior circulation stroke (79). Isolated transient neurologic episodes were 15 times more frequent in patients with subsequent vertebrobasilar stroke than infarcts in the carotid territory. Half of the transient isolated brainstem symptoms occurred within the 48 hours prior to the vertebrobasilar stroke, and 80% of patients with such transient neurologic episodes failed to seek immediate medical care at the time of the warning event.

Prevention

The greatest risk factor for stroke is age; with each decade, one's chance for stroke doubles. Arterial hypertension is, after age, the most important risk factor for stroke and transient ischemic attacks. Current stroke-prevention guidelines for patients with stroke or transient ischemic attack (and recommended in hypertensive patients with or without previous antihypertensive therapy) prescribe that treatment should be initiated or restarted at a level higher than 130/80 mm Hg several days after the transient ischemic attack. The optimum level for blood pressure reduction appears to be lower than 130/80 mm Hg according to office blood pressure levels (57). Up to now, no solid evidence exists showing the superiority of an antihypertensive agent in preventing vascular events in patients with transient ischemic attacks; it is thus advisable to make a personalized selection of the antihypertensive agent based on pharmacological properties, the mechanism of action, and the particular characteristics of the patient (57). Thiazide diuretics, angiotensin-converting enzyme inhibitors, or angiotensin II receptor blockers have been found to reduce recurrent stroke risk in randomized controlled trials and systematic reviews of randomized controlled trials (57). Evidence of the effectiveness of calcium channel blockers in reducing stroke recurrence is less conclusive than that of other blood pressure–lowering drugs; nonetheless, they may be used according to patient characteristics.

In addition to the pharmacological reduction of arterial hypertension, a thoroughgoing lifestyle modification is mandatory, comprising reduced sodium intake, weight loss, regular physical aerobic activity, limited alcohol consumption, and a healthy, balanced diet rich in fruits, vegetables, and low-fat dairy products (57).

Atrial fibrillation is an important risk factor for stroke, especially in the elderly. In the setting of primary stroke prevention, the recommended treatment for nonvalvular atrial fibrillation patients having a CHA2DS2-VASc score of 2 or greater in men or 3 or greater in women is oral anticoagulation with vitamin K antagonists (warfarin), direct thrombin inhibitor (dabigatran etexilate), and the factor Xa inhibitors rivaroxaban, apixaban, and edoxaban (44; 39).

According to the latest secondary stroke prevention guidelines from the American Heart Association, high-potency, high-dose statin therapy (daily atorvastatin 80 mg or rosuvastatin 20 mg) is the mainstay lipid-lowering therapy for individuals with a prior transient ischemic attack or ischemic stroke of noncardioembolic etiology without coronary heart disease and with a low-density lipoprotein (LDL) cholesterol level greater than 100 mg/dL. A high-potency, high-dose statin therapy is also recommended for patients with prior transient ischemic attacks or ischemic stroke and atherosclerotic disease; ezetimibe may be added when the LDL cholesterol level remains higher than 70 mg/dL, despite statin therapy. For individuals at very high risk of atherosclerotic cardiovascular disease whereby a target LDL cholesterol level is not achievable despite combined statin and ezetimibe therapy, addition of a proprotein convertase subtilisin/kexin type 9 (PCSK9) inhibitor is advisable (57).

Glucose metabolism disorders should be ruled out by measuring fasting plasma glucose, glycated hemoglobin (HbA1c), or an oral glucose tolerance test in all patients who have had a transient ischemic attack. Although the optimal target for glycemic control is still uncertain, the HbA1c goal should be individualized based on age, disease duration, macro- and microvascular complications, hypoglycemia risk, life expectancy, comorbidities, and patient preference. A reasonable goal for the long-term prevention of vascular events in patients with diabetic transient ischemic attacks is 7% to 8% of HbA1c. Glucose-lowering therapies with proven cardiovascular benefit are preferable (American Diabetes Association 2021; 57; 28).

Adherence to evidence-based stroke preventive strategies has been examined in an observational study of nearly 860,000 patients admitted to the hospital between 2007 and 2011 with brain ischemia (30% with a transient ischemic attack diagnosis) in the Get with the Guidelines (GWTG) Stroke Registry (06). Transient ischemic attack patients were more frequently treated in nonacademic hospitals. At hospital discharge, effective secondary prevention measures including antithrombotics, anticoagulation for atrial fibrillation, statin therapy, cessation of smoking, stroke education, and weight-loss recommendations if body mass index was greater than 25 kg/m2 were less prescribed in patients with a transient ischemic attack diagnosis compared to patients with a cerebral infarct diagnosis. With regard to antihypertensive therapy, no differences were observed in either of the patient cohorts. A trend over time to increase adherence to evidence-based preventive measures was observed in both groups of patients; nevertheless, this tendency was lower in patients with transient ischemic attacks (06).

Differential diagnosis

Confusing conditions

Misdiagnosis of acute ischemic stroke in the emergency setting is not uncommon. In a retrospective study including an academic and a community hospital, a quarter of 465 patients were misdiagnosed as stroke. The mean age of the patients with a stroke misdiagnosis was 71.5 years; 44% of these patients arrived at the hospital within the first 6 hours of stroke onset (05). Stroke misdiagnosis was more frequent at the community hospital (26%) than the academic hospital (20%). Misdiagnosis of stroke was nearly 2.5 times higher in patients with posterior circulation rather than anterior circulation stroke (37% vs. 16%). Dizziness, nausea, and vomiting—symptoms often due to posterior circulation stroke—favored stroke misdiagnosis (05).

In another retrospective study, stroke misdiagnosis occurred in 10.3% (156) of 1514 patients with a diagnosis of stroke; 141 patients had an ischemic stroke, and the rest had a cerebral hemorrhage. The mean age of patients with a stroke misdiagnosis was 77 years, and stroke misdiagnosis was more frequent when the vertebrobasilar territory was involved. Symptoms occurring in posterior circulation stroke, such as dizziness, nausea and vomiting, gait disturbance, and vertigo were more often associated with stroke misdiagnosis (99).

In the last few years, a meta-analysis of cerebrovascular event misdiagnosis in the emergency setting showed that stroke misdiagnosis is more likely when symptoms are transient, mild, and nonspecific, as in transient ischemic attacks. In fact, in three studies comprising 560 patients with transient ischemic attacks, misdiagnosis occurred in 41% of cases. In concordance with other studies, dizziness—a common symptom in vertebrobasilar ischemia—increases the likelihood of stroke misdiagnosis (95). Migraine, seizures, peripheral vertigo, and syncope are common transient cerebral ischemia mimics.

The relationship between migraine and ischemic stroke is complex. Migraine with aura is a stroke risk factor; migraine may cause a cerebral infarct, and a stroke may trigger a migraine attack. Moreover, aura symptoms often mimic focal neurologic deficits, and patients may, though uncommonly, experience typical aura symptoms with no headache (acephalgic migraine). Therefore, differentiating migraine with aura from a transient ischemic attack is challenging, both for emergency physicians and neurologists. A cohort study of 96 patients evaluated in the emergency setting described the clinical characteristics of 32 patients with migraine and aura, 32 patients with migraine aura without headache, and 32 patients with a transient ischemic attack diagnosis (30). In this study, patients with migraine and aura and those with migraine aura without headache were younger (mean age range 40 to 50 years) and more often women, compared to patients with transient ischemic attacks (mean age 65 years), who were more often men. Cardiovascular risk factors were more common in patients with transient ischemic attacks. Visual symptoms of aura occurred in 62% and 40% of patients with migraine aura without headache and migraine with aura, respectively, whereas 9% of patients with transient ischemic attacks had visual symptoms. Visual symptoms in aura may gradually spread from positive (scintillating scotomas, fortification spectra, and zigzag lines) to negative (hemianopsia and negative scotoma), likely due to cortical spreading depression, and such an evolution is quite distinct from the sudden onset of a transient ischemic attack (77). Preceding episodes of transient neurologic deficits were more common in patients with migraine aura without headache (81%) and migraine with aura (63%), whereas only 6% of patients with transient ischemic attacks had previous episodes. Two-thirds of episodes in patients with transient ischemic attack lasted less than 60 minutes; conversely, in around half of patients with migraine, episodes lasted from 1 to 24 hours.

Seizures manifest as recurrent stereotyped episodes. Unlike transient ischemic attacks, focal seizures often produce positive phenomena (such as elementary or complex visual hallucinations and paresthesias) that may migrate; postictal negative phenomena after focal seizures last around 2 to 3 minutes. Differentiation of generalized tonic-clonic seizures without focal onset from transient ischemic attacks rarely presents a diagnostic challenge; loss of consciousness rarely occurs in a transient ischemic attack, whereas lateral tongue biting, urinary incontinence, postictal confusion, fatigue, and myalgias usually follow a generalized seizure, as may postictal negative symptoms lasting up to 20 minutes (09). In elderly individuals, focal seizures with impaired awareness are common, postictal confusion may last longer than usual, and automatisms may be less common (58). Sometimes a reliable witness account is essential to defining ictal semiology. Elderly patients with new-onset epilepsy may have both cardiovascular risk factors and structural lesions causing seizures, such as older stroke, dementia, traumatic brain injury, and brain tumors.

In very exceptional cases, isolated, sudden, and transient amnesia may be the outstanding manifestation of a transient ischemic attack or ischemic stroke. Among the various acute-onset amnesic syndromes, transient global amnesia may be thought of as a transient ischemic attack in the acute clinical setting, considering that it affects middle-aged people with cardiovascular risk factors. Transient global amnesia is characterized by the sudden onset of dramatic anterograde amnesia lasting up to 24 hours and a variable long-term memory deficit with no altered consciousness or cognitive impairment. In a retrospective observational series of 390 patients with a clinical diagnosis of transient global amnesia (mean age 67 years, 60% women), 4% had had previous migraine attacks, whereas most patients had a single episode of transient amnesia (93). Mean episode duration was 3.2 hours in patients who arrived at the emergency department still showing amnestic symptoms; 61% of patients had a potentially relevant precipitating condition involving either physical or emotional stress. Small and single DWI-positive hippocampal lesions—40% located on the left hippocampus, 30% on the right, and the rest bilateral—were observed in 70% of patients. Eleven patients also exhibited punctuate lesions beyond the hippocampus on diffusion-weighted imaging; in four of these patients, such lesions were restricted to the posterior circulation. In all 11 patients, an extensive cardiac investigation was negative for a cardiac source of embolism (93). The precise pathophysiology of transient global amnesia is not yet well understood, but most of these attacks are not considered to be due to any of the common mechanisms associated with vertebrobasilar ischemia. In individuals with transient global amnesia (TGA), high-resolution brain MRI may reveal more small DWI hippocampal lesions than 1.5 Tesla MRI. At least one acute punctate DWI lesion was found in the hippocampal CA1 subfield in 56% of patients (71 in 126), mean aged 66 years with a clinical diagnosis of transient global amnesia and 3T brain MRI performed within 72 hours of symptom onset; unilateral hippocampal lesions were found to be more common. Additional acute DWI lesions outside of the hippocampus—mainly in the anterior circulation—were observed in 10% of the patients (33).

In another study, 7T brain MRI provided better detection of lesions on DWI sequences than 1.5/3T brain MRI in patients with a clinical diagnosis of transient global amnesia. Thirteen patients had both 1.5/3T and 7T brain MRI examinations. At least one acute punctate area of restricted diffusion on DWI sequences was visible in 11 (85%) of these 13 patients on 7T MRI and in six patients (46%) on 1.5/3T MRI. Moreover, 19 hippocampal lesions were found on 7T MRI, whereas 13 hippocampal lesions were observed on 1.5/3T MRI. Hippocampal lesions were easier to categorize as true findings on 7T MRI. Interestingly, no extrahippocampal lesions were detected (98).

In peripheral vestibular disease, patients suffer vertigo, occasionally associated with tinnitus and hearing loss, but symptoms are more prolonged than with transient ischemic attack. No other long tract, cranial nerve, or visual symptoms occur in patients with nonvascular disease of the vestibular apparatus. Only rarely is isolated vertigo, particularly if present for weeks, a manifestation of ischemia. Such patients often have cerebellar infarction, whereby vertigo and imbalance may be the only symptoms. Physicians must maintain a high index of suspicion in patients with vascular risk factors and vertigo (66). In a large multicenter prospective study, patients presenting with isolated transient dizziness in the emergency setting had a low risk of both subsequent stroke and transient ischemic attack at 2, 7, 30, and 90 days. In this study, 408 (4.2%) out of 11,507 patients diagnosed with transient ischemic attack (time-based definition) had transient dizziness with no other neurologic symptoms or signs. In both groups, neurologic symptoms lasted over one hour in more than 50% of patients. Isolated transient dizziness patients were older and had significantly more arterial hypertension, hypercholesterolemia, and coronary artery disease than their counterparts. Ninety percent of all patients had a brain CT scan in the emergency setting, whereas 10% had a brain MRI. Compared with patients with nonisolated dizziness, those with isolated transient dizziness had a significantly lower risk of cerebral infarct within 90 days (0.8% vs. 2.9%) and a lower risk of transient ischemic attack recurrence (1.7% vs. 5.6%). In the group of patients with isolated transient dizziness, the relative risk of cerebral infarct at 2, 7, 30, and 90 days was 0.18, 0.12, 0.26, and 0.29, respectively (07).

Syncope is rarely due to vertebrobasilar ischemia. When consciousness is lost in a vertebrobasilar transient ischemic attack, it usually lasts longer than syncope, and other vertebrobasilar symptoms or signs of focal deficit are present. Syncopal symptoms occur at transient ischemic attack or stroke onset in 5% of patients with no significant intracranial stenosis (82).

Transient focal neurologic episodes reminiscent of recurrent transient ischemic attacks may occasionally be a presenting feature in individuals with cardiovascular risk factors aged 65 years or older who later receive a diagnosis of cerebral amyloid angiopathy based on clinical criteria and blood-sensitive MRI neuroimaging sequences. In this setting, a focal subarachnoid hemorrhage in the convexity, often affecting the central sulcus, may manifest as short-lasting recurrent stereotyped episodes (amyloid spells), including positive (paresthesias, visual disturbance) or negative (weakness or speech impairment) symptoms. Focal paresthesias can rapidly migrate from the distal to the proximal end of a limb and may involve either the upper or lower limb or the face in various combinations. Although transient ischemic attacks may recur and usually manifest with negative focal deficits, a stereotyped pattern and migrating focal paresthesias are extremely unusual. The exact pathophysiology of such episodes is not well known, but it is thought that blood products trapped within a cerebral sulcus might produce recurrent waves of cortical depolarization or cortical spreading depression (92).

Structural and metabolic lesions may produce symptoms similar to transient ischemic attack symptoms by an unclear mechanism. Such lesions include tumors (especially with acute hemorrhage), subdural hematomas, aneurysms, amyloid angiopathy, multiple sclerosis, hypertensive encephalopathy, hypoglycemia, hyperventilation, and even primary or arteriovenous malformation-related intracerebral hemorrhage.

Diagnostic workup

Patients with transient ischemic attack should preferably undergo neuroimaging evaluation within 24 hours of symptom onset. Brain MRI should include T2-weighted sequences and, particularly, diffusion-weighted imaging as the preferred diagnostic imaging method (45; 89; 104; 31).). If MRI is not available, then head CT should be performed. Diffusion-weighted imaging may also be useful in differentiating between patients with transient ischemic attacks and those with other causes of transient neurologic symptoms.

Magnetic resonance diffusion-weighted imaging is highly sensitive in detecting ischemic tissue injury before changes are detectable on brain CT and standard MRI. DWI positivity was assessed in 47 studies (18 prospective, 17 retrospective, and of unknown design in the rest) covering a 17-year period from 1995 to 2012 that involved transient ischemic attack patients conforming to the classic time-based definition of a transient ischemic attack and excluding transient ischemic attack mimics. A positivity of 34% was found in diffusion-weighted imaging on MRI in 9000 transient ischemic attack patients; studies including larger patient cohorts reported lower DWI positivity than those with smaller cohorts. Thus, up to a third of patients with a transient ischemic attack diagnosis established by a neurologist or stroke specialist have negative DWI findings (08). Nevertheless, the rate of DWI-positive findings in patients with vertebrobasilar transient ischemic attacks may be lower than that reported in transient ischemic attack patients in general. In a small cohort of 40 posterior transient ischemic attack patients, only 12% had DWI-positive findings and were primarily associated with vertebral artery hypoplasia (64). In this particular patient group, the low frequency of DWI abnormality may have been due to compensatory mechanisms.

In the last few months, the rate of positive DWI magnetic resonance imaging was assessed, together with associated predictors, in 430 Chinese patients with a clinical diagnosis of transient ischemic attack, irrespective of the territory involved (105). Approximately one-third of these patients (126 out of 430) had a high-intensity lesion on DWI-MRI. The mean age was 61 years; 39% of patients reported a transient ischemic attack duration of less than 10 minutes, and in 42% the transient ischemic attack lasted between 10 and 59 minutes. The mean ABCD2, ABCD3, and ABCD3-I scores were 4, 6, and 8 points, respectively. The most common etiologies, based on TOAST criteria, were large- and small-artery disease, amounting to 38% and 32%, respectively. Nearly two-thirds of patients with high-signal intensity on DWI underwent dual antiplatelet therapy. Patients with motor weakness or speech abnormalities at admission had a 4-fold higher relative risk of abnormal DWI imaging; this risk was 13 times higher in patients with a high ABCD3-I score (105). Fifty-four percent of patients (240 out of 446) with transient ischemic attack (time-based definition) showed a high-intensity lesion on DWI. All patients underwent high-resolution MRI either at admission or at follow-up. Transient ischemic attack symptoms lasted over an hour in 62% of patients, and the mean ABCD2 score was 5 points. Perfusion deficits at admission and MRA-identified occlusion of the appropriate vessel were associated with DWI positivity (41).

Diffusion-weighted lesion status may be relevant for predicting stroke recurrence in the short aftermath of transient ischemic attacks. In certain cases, MRI perfusion imaging may be performed in the clinical scenario of a strongly suspected transient ischemic attack and normal DWI findings. In this setting, it is probable that the degree of hemodynamic compromise has not reached the tissue injury threshold. Focal perfusion deficits in the brain area corresponding to the symptoms are seen in about 25% of patients with clinical transient ischemic attack and negative DWI (35). Moreover, combining MRI-DWI and arterial spin labeling perfusion enables identifying more perfusion deficits in patients with transient ischemic attack than with DWI sequences alone (38).

In a study, DWI positivity was found in 88 (13%) of 633 transient ischemic attack patients; the median time between the index event and brain MRI was two days, and one-third of transient ischemic attacks were cryptogenic according to TOAST criteria (42). Compared to patients with transient ischemic attack and negative DWI, those with positive DWI had a 2-fold higher risk of recurrent ischemic stroke within 90 days after the index event, and the risk was nearly four times higher in patients with a cryptogenic index event.

Nearly one-half of patients with a clinical transient ischemic attack show no diffusion-weighted imaging evidence of cerebral ischemia. Early diffusion kurtosis imaging (DKI)—an extension of conventional diffusion-based MR methods—has been used along with DWI to predict stroke recurrence within 90 days in patients presenting an index transient ischemic attack (108). This study included 98 transient ischemic attack patients with no abnormalities on baseline DWI. Their mean age was 61.25 years and mean ABCD2 score was 4.12 points; their mean time from symptoms onset to MRI was 293 minutes and mean duration of symptoms was 27.2 minutes. A third of the patients with initial negative DWI showed hyperintensity on early DKI, and in most the location of this abnormality properly matched their clinical manifestations. No mention was made regarding the territory where the transient ischemic attacks occurred. During the 90-day follow-up period, a subsequent cerebral infarct was observed on DWI in 14 of the transient ischemic attack patients, 78% thereof within the first seven days after the initial transient ischemic attack. Two thirds of transient ischemic attack patients who had an abnormal early DKI developed a new cerebral infarct, and nearly all subsequent infarcts involved the same cerebral artery territory where the abnormality was observed on the initial DKI (108). This emerging neuroimaging method may, therefore, be a useful tool in predicting the risk of stroke subsequent to a transient ischemic attack.

Testing should exclude steno-occlusive disease of the vertebrobasilar arteries, cardiac disease, or hematological disease, in that order. Currently, imaging of the posterior circulation may be accomplished in most patients through noninvasive methods (84; 104). Spiral computed tomographic angiography is a noninvasive procedure to assess both extracranial and intracranial cerebral circulation. It may be a helpful tool for assessing the severity of intracranial stenosis and provides information about collateral circulation as well. Computed tomographic angiography has proven to be efficient at revealing steno-occlusive lesions in the anterior circulation, though it is less sensitive in detecting stenosis in the vertebrobasilar circulation. Computed tomographic angiography sensitivity and specificity was 97% and 78%, respectively, in the detection of large-vessel occlusion in the posterior circulation (63).

Contrast-enhanced MR angiography is an alternative to computed tomographic angiography, particularly when iodinated contrast is contraindicated. Moreover, in comparison with computed tomographic angiography, contrast-enhanced MR angiography had a slightly higher sensitivity and specificity in identifying vertebral artery stenosis greater than 50% (54).

Color duplex ultrasound can image the proximal vertebral arteries with reasonable accuracy. Conventional transcranial Doppler sonography for detection of significant stenosis in the distal vertebral or basilar arteries has a sensitivity ranging from 50% to 80% and a specificity of between 80% and 96% (37). This noninvasive approach enables full visualization of the proximal two thirds of the basilar artery and, though rarely, of its distal one third. Moreover, it is operator dependent, and an insufficient acoustic window is present in around 10% of patients.

The accepted “gold standard” for evaluation of arterial occlusive disease has traditionally been catheter angiography. This invasive technique, however, is time-consuming and has substantial risks. Intra-arterial digital subtraction angiography should be reserved for those in whom intra-arterial therapeutic interventions are planned or for whom noninvasive tests are not diagnostic.

In patients with normal or mildly diseased vertebrobasilar arteries, or in patients at high risk of embolic cardiac disease, cardiac testing is initiated. When a cardiac source of embolism is suspected or to be ruled out, both transthoracic and transesophageal echocardiography are useful stroke diagnostic tools. The latter is more sensitive than transthoracic echo and, thus, often the preferred cardiac imaging modality. To date, cardiac MRI and CT are considered complementary to echocardiography in selected cardiac sources of embolism (10).

The yield of echocardiography was assessed in 869 acute transient ischemic attack patients with a mean age of 66 years whose median ABCD2 score was 4 points (102). The ABCD2 score for half of these patients was between 4 and 5 points, although 15% of them scored between 6 and 7. In this cohort, 70% (603) of the patients underwent echocardiography (mostly transthoracic, although 10% of them had either transesophageal or both transthoracic and transesophageal echocardiography). Ten percent of the transient ischemic attack patients undergoing echocardiography had identified sources of cardiac embolism, the most common of which were a complex aortic arch atheroma, patent foramen ovale, apical akinesis, and an ejection fraction of less than 35%. Transesophageal echocardiography identified a source of cardiac embolism in 31% of patients, and transthoracic echocardiography did so in 8% of patients. Holter monitoring is performed when an emboligenic arrhythmia such as intermittent atrial fibrillation is suspected.

In patients for whom vertebrobasilar imaging and cardiac tests are unrevealing, vasculitis tests or specialized coagulation tests (antiphospholipid antibodies and other hypercoagulable states) can be considered, particularly in young patients without vascular risk factors.

Management

Transient ischemic attack should be treated as a medical emergency, with prompt and appropriate investigation to determine the mechanism of ischemia and the subsequent acute and long-term therapy to be administered. Although there is ongoing debate as to the appropriate organization model to facilitate rapid assessment and prompt, proper treatment for patients with transient ischemic attacks, hospitalization is still advised for patients with transient ischemic attack within the previous 48 hours or with crescendo (multiple and increasingly frequent) transient ischemic attacks (45). Therefore, outpatient management should be mainly reserved to patients with an estimated low risk of early stroke after a transient ischemic attack. In this setting, several strategies have been developed for expeditious assessment and treatment of such patients. In general terms, institutions providing care for patients with transient ischemic attacks need to have an organized system for evaluation and management that should be individualized on the basis of care settings, resources, and patient characteristics. For example, a triage protocol in the emergency department for patients with suspected transient ischemic attack allowed selection of patients for hospital admission based on the ABCD2 clinical score in combination with early cervical and intracranial vessel imaging (76). In this protocol, two thirds of 224 patients with suspected transient ischemic attack seen initially in the emergency department were discharged to complete their evaluation in a specialized outpatient setting. The recurrent stroke rate was lower in the group of patients referred to the transient ischemic attack clinic than in the hospitalized patients, suggesting that patients with low risk of stroke can be reliably evaluated and managed in an outpatient setting.

All patients should undergo modification of their cardiovascular risk factors, including reduction of blood pressure, rigorous glucose control, cessation of smoking, reduction in alcohol consumption in heavy drinkers, and weight reduction.

In daily clinical practice, patients with noncardioembolic transient ischemic attack should be immediately started on long-term daily antiplatelet therapy. Aspirin monotherapy in doses ranging from 50 to 325 mg/day or a dual antiplatelet therapy (a combination of 200 mg twice daily of extended-release dipyridamole plus aspirin 50 mg/day) are options for initial therapy (57). Clopidogrel 75 mg/daily is also an acceptable antiplatelet agent; it also applies in cases of aspirin allergy. Selection of an antiplatelet agent should be individualized on the basis of patient risk factor profiles, cost, tolerance, and other clinical characteristics. Short-term dual antiplatelet therapy (acetylsalicylic acid in combination with clopidogrel) is currently recommended over single antiplatelet therapy after a noncardioembolic high-risk transient ischemic attack (ABCD2 score of at least 4 points) and is followed by long-term antiplatelet monotherapy. Dual antiplatelet therapy should be initiated as soon as possible within the first seven days after a transient ischemic attack and continued for anywhere from 21 to 90 days (31; 57). The updated guidelines for secondary stroke prevention from the American Heart Association suggest 30 days of dual antiplatelet therapy (ticagrelor plus acetylsalicylic acid) given within the first 24 hours after a transient ischemic attack in patients with an ABCD2 score of 6 or greater or symptomatic intracranial or extracranial stenosis (narrowing of 30% of more in the lumen diameter of an artery that could account for the transient ischemic attack) to reduce the risk of early recurrent stroke, although there is a concomitant increased risk of hemorrhagic complications, including intracranial bleeding (57).

In a meta-analysis of randomized clinical trials from 2005 to 2014 in patients with transient ischemic attack or cerebral infarction who were allocated either to dual antiplatelet or single antiplatelet therapy, the data from five trials showed that dual antiplatelet therapy (mainly aspirin plus clopidogrel) in 2984 patients with transient ischemic attack reduced the risk of recurrent ischemia by around 20%, although not significantly, compared to 2972 patients with transient ischemic attack who were placed on antiplatelet monotherapy (65).

Dual antiplatelet therapy reduced the risk of major ischemic events (ischemic stroke, myocardial infarction, or death from vascular origin) by 25% at 90 days after index event in patients with high-risk transient ischemic attack and minor ischemic stroke compared with patients treated with mono antiplatelet therapy (48). Moreover, a 28% reduction in ischemic stroke risk occurred in patients on dual antiplatelet therapy compared with those on single antiplatelet drug treatment. Nevertheless, patients treated with a combination of two antiplatelets had a significant 2-fold increase in risk of major hemorrhagic complications (mainly nonfatal nonintracranial hemorrhages) compared with patients treated with a single antiplatelet agent. In this double-blind placebo-controlled trial, 2108 patients had a transient ischemic attack (ABCD2 score higher than 4 points) and 2773 patients had a minor ischemic stroke (3 or fewer points on the NHISS scale). Patients were randomly allocated to combined antiplatelet therapy or a single antiplatelet agent plus placebo within 12 hours after the qualifying event. Patients on dual antiplatelet therapy received a 600 mg loading dose of clopidogrel and then 75 mg of clopidogrel per day plus an open-label dose of aspirin ranging between 50 and 325 mg from day 2 to day 90, whereas patients on monotherapy received aspirin (50 to 325 mg per day) plus a placebo. The median ABCD2 score in the 2108 patients with transient ischemic attack was 5 points. Patients who had an ABCD2 score higher than 5 points showed close to a 50% reduction in major ischemic event risk; however, the number of patients in both antiplatelet therapy groups was very low. When patients with a score of 5 points or lower had a transient ischemic attack as the qualifying event, there were no differences in the risk of major ischemic events between the groups (48).

A pooled analysis of the POINT and CHANCE trials – large-scale, randomized, double-blind, placebo-controlled clinical trials –showed that early and short course of dual antiplatelet therapy provide greater protection against further ischemic stroke than antiplatelet monotherapy in patients with a transient ischemic attack or nondisabling ischemic stroke of noncardioembolic etiology (78). The Platelet-Oriented Inhibition in New TIA and Minor Ischemic Stroke (POINT) trial included 4881 patients, and the Clopidogrel in High-Risk Patients with Acute Nondisabling Cerebrovascular Events (CHANCE) trial enrolled 5170 Chinese patients. In both trials, patients with minor stroke and transient ischemic attack were randomized to dual antiplatelet therapy (clopidogrel plus aspirin) or aspirin therapy alone (for 90 days in the POINT trial and 21 days in the CHANCE trial). Patients who underwent dual antiplatelet therapy received a loading dose of clopidogrel (600 mg in the POINT trial and 300 mg in the CHANCE trial) followed by a daily dose of 75 mg. The daily dose of aspirin ranged between 50 and 325 mg in the POINT trial and was 75 mg in the CHANCE trial. Antiplatelet therapy was started within 12 hours (POINT) and 24 hours (CHANCE) after the qualifying event. Patients in both trials were followed for 90 days. The pooled data of these trials showed that 5016 patients were randomized to dual antiplatelet therapy and 5035 patients to aspirin alone. Minor stroke (65%) and transient ischemic attack (35%) were the qualifying events. Compared to aspirin monotherapy, clopidogrel plus aspirin significantly reduced the risk of ischemic stroke, the composite of ischemic stroke, disabling or fatal stroke, myocardial infarction, and death from ischemic vascular origin by one-third. Dual antiplatelet therapy provided a clinical benefit within 90 days, particularly in the first 21 days of treatment, but the risk of hemorrhagic adverse outcome was greater, though not significantly, compared with antiplatelet monotherapy (78).

Another dual-antiplatelet therapy trial assessed whether a ticagrelor and aspirin combination is superior to aspirin monotherapy in reducing the 30-day risk of stroke recurrence in high-risk (ABCD2 6 or higher) transient ischemic attack patients and patients with nondisabling, noncardioembolic ischemic stroke (NIHSS ≤ 5) (47). Patients in this double-blind trial were randomly allocated within the first 24 hours after the qualifying event to either ticagrelor plus aspirin (5523 patients) or placebo plus aspirin (5493 patients) administered over 30 days. The ticagrelor group received a loading dose of 180 mg followed by 90 mg twice daily; 300 to 325 mg of aspirin was given on the first day, followed by 75 to 100 mg daily in both arms of the study. In this trial, the risk of recurrent stroke or death was significantly lower in the ticagrelor-aspirin group; however, dual antiplatelet therapy resulted in a significant increase in severe bleeding. A transient ischemic attack was the qualifying event in 491 patients in the ticagrelor-aspirin group and in 540 patients in the aspirin group. A 20% relative risk reduction in the primary outcome (recurrent stroke or death) was observed within 30 days in the ticagrelor group compared to the aspirin group (47). Based on a new definition of efficacy and safety endpoints, this double-blind randomized trial of dual antiplatelet therapy showed a net clinical benefit that outweighed the hemorrhagic risk. In terms of absolute risk reduction transformed into patient numbers, in a cohort of 1000 patients treated with ticagrelor-aspirin therapy for 30 days as opposed to aspirin alone, there would be an estimated reduction of 12 major ischemic events (composite of ischemic stroke and nonhemorrhagic death) and an additional three major hemorrhage events (composite of intracerebral hemorrhage and fatal bleeding) (46).

Long-term oral anticoagulation, either with vitamin K antagonists (warfarin) or nonvitamin K antagonists (also called new oral anticoagulants) is recommended for treating patients with transient ischemic attacks due to paroxysmal, persistent, or permanent nonvalvular atrial fibrillation. The latest American Heart Association guidance for secondary stroke prevention recommends early anticoagulation with new oral anticoagulants (direct thrombin inhibitors or factor Xa inhibitors) over vitamin K antagonists for patients with atrial fibrillation and transient ischemic attack who do not have moderate to severe mitral stenosis or a mechanical heart valve (57).

Vitamin K antagonists are recommended to reduce stroke risk in patients with transient ischemic attack with atrial fibrillation and severe mitral stenosis or a mechanical heart valve. Moreover, long-term anticoagulation, preferably with vitamin K antagonists, is recommended for patients with a transient ischemic attack and mitral stenosis or a mechanical heart valve prothesis. At least three months of vitamin K antagonist anticoagulation therapy is recommended for patients with a transient ischemic attack and a history of recent myocardial infarction with reduced ejection fraction, for patients with left ventricular thrombus, and for patients with left atrial or left appendage thrombus in the setting of ischemic, nonischemic, or restrictive cardiomyopathy. Usually, a 2.5 (range 2.0 to 3.0) INR target is recommended when vitamin K antagonists are administered (57).

Patients with posterior circulation transient ischemic attacks due to extracranial vertebral stenosis may be treated surgically or by the endovascular approach, although their usefulness is still uncertain and reserved for patients with recurrent symptoms refractory to medical therapy (57). Potential surgical treatments include arterial reimplantation, extracranial to intracranial anastomosis, surgical endarterectomy, and endovascular angioplasty with or without stenting. Revascularization procedures (extra- or intracranial) are influenced by location of stenosis or occlusion, patients’ symptoms, and the surgeon’s experience. Extracranial reconstructions include transposition of the vertebral artery on the common carotid artery (proximal vertebral artery stenosis) and vertebral to carotid bypass (vertebral artery stenosis distal to the origin). Intracranial revascularization procedures include several bypasses: the superficial temporal artery, occipital artery or external carotid artery to the superior cerebellar artery, posterior cerebral artery, posterior inferior cerebellar artery, or anterior inferior cerebellar artery (16). There is no good evidence that endarterectomy in the posterior circulation is useful.

As mentioned above, percutaneous transluminal angioplasty with or without stenting of the vertebral and basilar arteries may be performed to treat patients with severe stenoses and recurrent ischemic symptoms despite the best medical therapy. Newer percutaneous transluminal angioplasty and stent devices will likely be more streamlined and flexible, permitting safer and more reliable delivery into the cerebrovascular system. Percutaneous angioplasty with stent placement in 33 patients with extracranial vertebral stenosis (including 21 patients with vertebrobasilar transient ischemic attacks alone or prior to posterior circulation infarct), showed high levels of technical success in improving the angiographic appearance of the stenosis and a low rate of periprocedural complications and deaths related to posterior circulation disease. A substantial improvement of ischemic symptoms was also observed in most patients. Medium-term follow-up, however, disclosed a high rate of moderate to severe recurrent stenosis although most patients remained asymptomatic (01). A high rate of stent placement with a low rate of complications was reported in another cohort of 63 patients, which included 18 patients with extracranial vertebral stenosis (greater than 50%) and 23 patients with intracranial vertebral or basilar stenosis. At 30 days, most strokes occurred in patients with vertebrobasilar stenosis rather than patients with anterior circulation stenosis. At six months, restenosis (greater than 50%) was observed in 35% of cases, mainly in the extracranial vertebral arteries. Stroke occurred in 7% of patients (half of these in the posterior circulation) between 30 days and 1 year after stenting (91).

Some interesting results have been reported for intracranial stent placement in symptomatic anterior and posterior circulation stenosis (greater than 50% measured on digital subtraction angiography). Stenting in 165 cases of vertebral artery (V3 and V4 segments) and basilar artery stenoses showed a technical success rate of 91%. In-hospital disabling and nondisabling ischemic events were recorded in 4.8% and 6% of patients, respectively; a similar figure occurred for hemorrhagic complications and death (2.4% of cases each). When analyzing the results for anterior and posterior circulation procedures, it was found that technical failure and hemorrhagic complications were more frequent in anterior circulation stent placement, and ischemic events were slightly higher in posterior circulation stenting (59).

Moreover, the clinical benefit of vertebral stenting over exclusively medical therapy (antithrombotic, statins, and antihypertensive agents) was reported in 39 patients (70% with transient ischemic attacks) having severe symptomatic (less than 48 hours) ostial vertebral artery stenosis (51). Procedural success was achieved in all 10 patients undergoing vertebral stenting, and periprocedural complications (transient ischemic attack) occurred in one patient within 30 days after stenting. The restenosis rate among all these patients was 10%. At four years of follow-up, patients with vertebral stent placement had a lower risk of transient ischemic attack and stroke (fatal and nonfatal) in the posterior circulation than did medically treated patients (10% vs. 45%).

To assess the safety and efficacy of percutaneous transluminal angioplasty and stent placement in the treatment of symptomatic stenosis of intracranial arteries, the National Institute of Neurological Disorders and Stroke (NINDS), sponsored the Stenting vs. Aggressive Medical Management for Preventing Recurrent Stroke in Intracranial Stenosis (SAMMPRIS) study, a randomized phase III clinical trial, in November 2008. The trial compared the Wingspan self-expanding stent and Gateway PTA balloon catheter procedure plus intensive medical therapy against intensive medical management alone in patients with high-grade (angiographically proven 70% to 99%) stenosis of a major intracranial artery. All 451 of the patients in the study were enrolled within 30 days of a transient ischemic attack or nondisabling stroke attributed to intracranial arterial stenosis and were randomly assigned to a medical management group (227 patients) or a percutaneous transluminal angioplasty and stenting (PTAS) group (224 patients) (15). The medical management group included 22 patients with vertebral stenosis and 51 patients with basilar stenosis; the PTAS group had 38 vertebral stenosis and 49 basilar stenosis patients. The PTAS group showed a significantly higher rate of 30-day stroke and death (14%) compared to the medical management group (5.8%). Beyond the 30-day mark, the rates of stroke in the territory of the stenotic artery were similar in both groups, but fewer than half the patients were followed for 1 year (15).

Based on these unexpected results, the NIH-appointed Data and Safety Monitoring Board recommended the SAMMPRIS Executive Committee stop the prospective trial in April 2011 (15). The trial’s results give rise to a number of concerns, among which are the appropriateness of the PTAS group’s patient selection, operator expertise required to assure the benefits of this endovascular procedure, the proper timing of ischemic event evaluation in medically managed patients to avoid underestimating such events, and the real threshold of death and periprocedural complications of PTAS to be considered in future trials.

The SAMMPRIS trial reported on the frequency of symptomatic in-stent restenosis and the rates of stroke recurrence (either transient ischemic attack or symptomatic infarction) in the stented artery territory 30 days after endovascular treatment (24). Out of 17 patients with nonprocedural symptomatic infarction, six (35%) had infarcts in the posterior circulation (mostly nondisabling stroke, except in one case), and four had in-stent restenosis estimated by digital subtraction angiography between 4 and 28 months after enrollment. Moreover, in the stenting arm of this trial, 4 out of 26 patients (15%) had a cerebral infarction with temporary signs in the posterior circulation occurring between 1.5 and 27 months after enrollment. One patient had in-stent restenosis estimated by computed tomographic angiography (CTA). Vertebrobasilar transient ischemic attacks were recorded in 6 of 26 patients (23%) in the stented group between 3 and 18 months after enrollment (24).

The results of the Vertebral Artery Ischaemia Stenting Trial (VIST) were reported in the last few years (68). This prospective, open-label, blinded endpoint clinical trial randomized 182 patients with symptomatic intra- and extracranial atherosclerotic vertebral artery stenoses (over 50% determined by digital subtraction angiography) who had had either a transient ischemic attack (approximately 32%) or a nondisabling infarct (68%) within the 3 months prior to randomization. Patients were allocated to best medical therapy alone (medical group) or angioplasty with or without stenting plus best medical therapy (stent group). Most stenoses (79%) in the stent group were extracranial. In the stent group, all 61 patients were stented, but three also received angioplasty; two patients with intracranial stenosis had major complications (a fatal subarachnoid hemorrhage and a brainstem stroke), and a third patient with extracranial vertebral artery stenosis had a nonfatal stroke. During a mean follow-up of 3.5 years, “any stroke” (the primary endpoint) occurred in five patients (with one fatality) in the stent group and in 12 patients (with two fatalities) in the medical group. Thus, these results show an absolute risk reduction of 25 strokes per 1000 person-years in the stent group (hazard ratio of 0.40) compared to the medical group. Additionally, the hazard ratio for stroke and transient ischemic attack was 0.5 (p = 0.05). Three patients had a stent occlusion during follow-up (two were asymptomatic, and one had a transient ischemic attack) (68). Because the majority of patients in the trial had extracranial stenosis, this precluded drawing firm conclusions on the benefit of stenting in intracranial cases. This trial appears to suggest that the risk of periprocedural stroke is higher for intracranial stenoses than for extracranial stenosis.

Although knowledge, experience, and technological advances in endovascular stroke treatment have evolved over the last few years, the role of neurointerventional procedures in patients with posterior ischemic stroke—particularly in the setting of intracranial stenosis—needs to be appropriately defined in light of the trials mentioned above.

Special considerations

Pregnancy

Pregnancy may produce a hypercoagulable state leading to stroke, but these strokes are not commonly preceded by transient ischemic attacks.

Anesthesia

Head and neck trauma secondary to neck manipulation could occur during anesthesia, with consequent vertebrobasilar dissection, but this is rare. Also, in most patients with transient ischemic attack taking warfarin or aspirin, the medications can usually be safely stopped for a brief period for surgery.