Clinical indications and description

Supportive care for ischemic stroke patients has two main objectives: (1) to minimize injury to potentially ischemic brain tissue and (2) to prevent and treat the many neurologic and medical complications that may occur in the immediate period following stroke. As far as advancing therapeutics in ischemic and hemorrhagic stroke is concerned, it continues to have more broad potential of including those candidates who, at some point, were thought to have no interventions (55). Duncan and colleagues described the importance of post-acute care and outcome follow-up (23). They indicated the need for a paradigm shift regarding rehabilitation readiness of comprehensive stroke care centers where most stroke needs are met.

As cerebral blood flow drops below approximately 25 ml/100 gm tissue/min, neurons begin to malfunction, and neurologic symptoms become manifest, but normal function may be restored with reperfusion. As blood flow is reduced further, neurons begin to suffer irreversible damage. The extent of irreversible neuronal injury (ie, infarction) varies based on the degree and the duration of hypoperfusion (37). The region of cerebral tissue that is hypoperfused yet not irreversibly injured is referred to as the “ischemic penumbra.” Measures to increase cerebral perfusion may prevent permanent damage to this penumbral tissue. Furthermore, additional factors such as hyperglycemia, hyperthermia, and endothelial dysfunction may negatively impact salvage of the penumbra by accelerating neuronal damage. Many of the medical and neurologic complications to which stroke patients are susceptible may contribute to neuronal injury in an obvious manner as in the case of cerebral edema or hypoxemia, or more subtly as in the case of infection causing fever or the release of inflammatory cytokines (57). There is, therefore, considerable overlap between the goals of protecting neuronal tissue and preventing medical and neurologic complications. Saini and colleagues stated that out of approximately 14 million strokes in 2016, 10% to 20% had large vessel occlusion, less than 5% of patients received thrombolytics, and less than 1% received mechanical thrombectomy (77).

Indications

All ischemic stroke patients should receive supportive care. Although there may be a natural inclination to be less aggressive in patients with major stroke, particularly in older patients, it should be noted that even in these populations the majority of patients will survive their stroke. The degree of functional recovery, however, may be dramatically impacted by the intensity and appropriateness of supportive care.

Description

Blood pressure management. In healthy individuals, cerebral blood flow is held constant across a wide range of systemic blood pressures. In contrast, the cerebral vasculature in acute stroke patients is unable to adjust to variations in systemic blood pressure, and the relationship between cerebral blood flow and blood pressure becomes linear (64). This impairment in cerebral autoregulation suggests that lowering systemic blood pressure may decrease cerebral perfusion and increase ischemic brain injury. Both extremes of low and high blood pressure may be detrimental. Observational studies correlating initial blood pressure values and stroke outcome suggest a U-shaped curve (52). The theoretical concern that acutely lowering blood pressure could be harmful is supported by results from a small randomized trial testing nimodipine as a neuroprotectant agent; in the trial, a correlation between medication-induced blood pressure reduction and worse clinical outcome was seen (03). A small randomized controlled trial of candesartan, an angiotensin receptor blocker, in acute stroke did report decreased mortality at 12 months. However, no significant differences in blood pressure were reported (81). A small randomized trial of blood pressure reduction in acute stroke reported no early harm and borderline benefit at 3 months (70); however, both ischemic and hemorrhagic stroke patients were pooled together making interpretation difficult. A larger and more definitive trial of gradual blood pressure lowering again using candesartan in acute stroke showed no benefit from early blood pressure reduction and possible harm in a trial pooling ischemic and hemorrhagic stroke patients (79). Results of the INTERACT trials suggest blood pressure management in acute intracerebral and acute ischemic stroke should be considered separately; furthermore, aggressive blood pressure lowering in acute intracerebral hemorrhage may be beneficial (06; 05). A large trial of blood pressure lowering versus no antihypertensive therapy limited to acute ischemic stroke, which involves over 4000 patients, has been published (34). No advantage to blood pressure lowering could be demonstrated, even with this population where subjects with known large vessel stenosis were excluded.

As a general rule, therefore, blood pressure should not be actively lowered in patients with acute cerebrovascular ischemia. Exceptions include patients who are candidates for or have already received thrombolytic therapy because the risk of intracranial hemorrhage is increased in patients with severe hypertension (92), patients with evidence of active hypertensive injury to other organs (eg, myocardial ischemia), and patients with extremely severe hypertension (greater than 220/120). It is generally believed that decisions about blood pressure management should also take into consideration the patient's baseline blood pressure status.

Table 1 summarizes current recommendations from the American Stroke Association, including recommendations for patients eligible for thrombolytic therapy (02). In this latter group, the standard protocol for blood pressure management is based on the National Institute of Neurological Disorders and Stroke trial of tPA for acute stroke and should generally be rigidly adhered to.

Table 1. Guidelines for Treatment of Elevated Blood Pressure in Ischemic Stroke

Beyond avoiding therapies that lower blood pressure, induced hypertension using vasopressive agents has been suggested as a possible therapy for acute stroke. In a cohort of patients with acute hemispheric stroke, vasopressor-induced increases in mean arterial pressure were associated with increases in cerebral perfusion pressure (rising from 72.2+/-2 to 97+/-1 mm Hg, P< 0.0001) (85). Of note, there was also a significant, though modest, increase in ICP (rising from 11.6+/-0.9 to 11.8+/-0.9 mm Hg, P< 0.05). A small pilot trial of 13 patients treated with vasopressor-induced hypertension demonstrated increased perfusion and improved outcomes (35). However, a meta-analysis of 12 small studies concluded that the benefits and risks remain uncertain; thus, the use of vasopressors to induce hypertension must be considered experimental until benefit is demonstrated in randomized trials (59). Augmentation of cerebral perfusion by partial aortic occlusion is a related intervention that appears safe, but it awaits clinical confirmation (87). An improved understanding of dynamic and steady-state autoregulation, both acutely and in the longer term, is needed (07).

In all patients, hypotension should be aggressively treated and the underlying etiology (ie, myocardial infarction, sepsis) determined and addressed. Initial treatment should include rapid volume replacement with normal saline. If blood pressure does not improve, vasopressive agents should be used.

Volume status. Stroke patients are at a high risk of volume depletion due to decreased oral intake and increased insensible losses. Hypovolemia (based on serum osmolality) has been associated with worse outcomes and increased mortality in acute ischemic stroke (11). Volume depletion undoubtedly has many deleterious effects. One study of 104 patients suggests that dehydration may be progressive poststroke, and increased osmolality (greater than 297 mOsm/kg) confers a nearly 5-fold increased risk of venous thromboembolism in multivariate analysis (44). Isotonic saline, ie, “normal” or 0.9%, should be used for volume repletion and maintenance as hypotonic saline has the potential to exacerbate cerebral edema. Dextrose infusions should be avoided as hyperglycemia has been associated with worse outcomes following stroke.

Positioning in bed. In patients with middle cerebral artery occlusion, upright posture has been shown to decrease cerebral blood flow and increase oxygen extraction in the affected hemisphere (65). Also, small studies have demonstrated that positioning patients flat in bed can improve cerebral blood flow (84). In patients with acute middle cerebral artery stroke, lowering the head of the bed from 30 degrees to 15 degrees increased middle cerebral artery mean flow velocity measured by transcranial Doppler by 12% (p=0.001), and by an additional 8% when lowered from 15 degrees to 0 degrees (p=0.016) (99). Importantly, there was no change in blood pressure or heart rate with these changes in head position. Based on these data, it seems reasonable to maintain acute stroke patients in the flat or nearly flat position acutely. There are no data to suggest a definitive time period; however, a common practice is to maintain the flat position for the initial 24 hours after presentation. If neurologic status worsens on head elevation, the patient should be returned to the flat position, and measures to improve cerebral perfusion and maintain collateral blood flow should be considered. In patients with increased intracranial pressure due to swelling from large hemispheric stroke, elevating the head of the bed may improve venous drainage and reduce ICP, but this may come at the cost of reduced cerebral perfusion pressure. There are few data to guide decision making in this difficult situation. In patients who cannot tolerate lying flat in bed due to orthopnea or other medical conditions, the head of the bed should be kept at the lowest level tolerated by the patient. Finally, frequent changes in body position (regardless of head position) may help awake and alert patients tolerate lying flat and are indicated to minimize the risk of pressure sores. More data on aspiration risk, the risk of decreased venous outflow, and net clinical benefit are needed.

Glycemic control. Considerable evidence supports a link between hyperglycemia and poor outcome after stroke. One study suggests that capillary glucose levels higher that 155 mg/dL anytime in the first 48 hours after admission for acute stroke is associated with a 2.7-fold risk of poor outcome at 90 days after corrected for stroke severity, infarct volume, known diabetes, or age (29). Compared with normoglycemic patients, acute stroke patients who present with hyperglycemia have larger infarct volume, poorer functional outcome, and a higher likelihood of hemorrhage if treated with thrombolysis (24; 46; 92). Hyperglycemic patients also have reduced salvage of at-risk brain tissue within the ischemic penumbra (67; 75). Data in critically ill patients without stroke suggest a benefit to intensive glucose control, although controversy exists regarding desired stringency. In a large randomized trial of patients admitted to a surgical intensive care unit, strict normalization of serum glucose with an insulin drip reduced multiorgan failure, sepsis, and overall mortality (96). Mortality reduction has also been shown in a similar trial of patients with myocardial infarction (53). The NICE-SUGAR trial enrolled and randomized over 6000 ICU patients (medical or surgical) to either intensive hyperglycemic control (81 to 108 mg/dL) or less than 180 mg/dL (10 mmol/L) and found increased mortality at 90 days without evidence of benefit in the intensive control group (61).

At present, there are no data proving a benefit of glucose lowering in acute stroke patients. A prematurely closed, underpowered trial of aggressive glucose control employing continuous glucose-insulin-potassium infusion in over 900 patients was safe but failed to show benefit in terms of the primary endpoint, mortality at 90 days, or the secondary endpoint, mortality or poor outcome (33). A pilot study of a somewhat different design demonstrated safety and hinted at efficacy (41). A definitive trial remains to be done. At present, strict glucose control using frequent finger-stick glucose checks and aggressive sliding scale insulin are reasonable, regardless of whether the patient has a known history of diabetes. Studies of hyperacute stroke in animal models suggest insulin effects on glutamate, unrelated to glucose levels, may be beneficial (26).

Hypoglycemia is a well-known stroke mimic and may present with focal neurologic symptoms. Prolonged hypoglycemia may also directly injure neurons. Thus, serum glucose should be checked immediately on presentation, and hypoglycemia should be urgently treated with a dextrose infusion. Overall, both hypoglycemia and hyperglycemia are risk factors for worse outcomes in acute ischemic stroke by promoting hemorrhagic transformation. Intensive insulin therapy utilized for aggressive hyperglycemia management has increased rates of hypoglycemia, which may be associated with larger infarct growth (48).

Temperature control. Fever is a common finding in acute stroke patients, with as many as 25% generating a temperature of 38.0°C or more within the first 48 hours of admission (32). Increased core temperature in the acute stroke patient may result in an increase in neuronal metabolic demands, neurotransmitter release, inflammatory mediator activity, and free-radical production within the ischemic penumbra. Body temperature in acute stroke patients has been correlated with initial stroke severity, infarct size, mortality, and functional outcome in survivors (97). For each 1°C increase in body temperature, the risk of death or severe disability more than doubles (72). A large observational study found that 35% of ischemic stroke patients were febrile sometime during acute hospitalization, but, additionally, the study demonstrated a strong correlation between fever burden (peak temperature duration product) and dismal outcome, death, or hospice such that patients with a fever burden of more than four “degree days” had a nearly 7-fold risk of dismal outcome (69). Therapeutic hypothermia has been shown to improve neurologic outcomes in patients with cardiac arrest (09; 93). However, important differences between resuscitated cardiac arrest and stroke (global vs. focal brain insult, complete reperfusion vs. incomplete or no reperfusion) limit extrapolation of these trials. Numerous trials of induced hypothermia in stroke are underway. However, until data from these trials demonstrate a benefit of hypothermia, normothermia (temperature above 35.5°C and below 37.5°C) should be the goal. Lastly, it is widely held that fever in an acute stroke patient should be aggressively treated with antipyretics and possibly ice packs or other cooling devices, concurrent with a thorough search for an infectious source.

Oxygenation. One of the more intuitive priorities in the care of the acute stroke patient is the maintenance of adequate tissue oxygenation. However, given that many stroke patients do not have significant lung disease or respiratory compromise, it is not clear that all patients require supplemental oxygen. A quasi-randomized study found no benefit of supplemental oxygen given to stroke patients (73). A randomized trial of high flow oxygen via face mask in patients with ischemic penumbra present on MRI diffusion and perfusion imaging reported reduced diffusion volume at 4 hours and improved National Institutes of Stroke Scales at 1 week, but there were no significant differences at 3 months (88). These investigators also showed strong correlation between favorable apparent diffusion coefficient changes and MR spectroscopy measurements of brain lactate and n-acetyl-aspartate (89); however, clinical utility remains to be demonstrated. American Stroke Association recommendations call for supplemental oxygen to be given as needed to maintain an oxygen saturation of more than 92% by pulse oximetry or blood gas (02). There is no convincing evidence to suggest benefit from hyperbaric oxygen in stroke, with the exception of stroke due to arterial air embolism (04; 02).

Statin therapy. Statins (3-hydroxy-3-methyl-glutaryl-CoA reductase inhibitors) are now recommended for secondary ischemic stroke prevention in patients with low-density lipoprotein (LDL) greater than 100 mg/dl or greater than 70 mg/dL in patients with diabetes (76). In-hospital initiation of statin therapy poststroke is encouraged and may be advantageous in terms of compliance (80). Discontinuation of statin therapy initiated previously on hospitalization for acute ischemic stroke may worsen outcomes (14), whereas continuing statin therapy or initiation of statin therapy in acute ischemic stroke may be beneficial (12; 28; 62). Although formal guidelines have yet to be formulated and larger studies of acute or hyperacute statin therapy are needed, the continuation of statin therapy and very early initiation of statin therapy in ischemic stroke patients are generally supported by the available data and now common practice (66).

Prevention and management of medical and neurologic complications

Prevention and management of medical complications. Medical complications such as pneumonia, sepsis, pulmonary embolism, and myocardial infarction account for approximately half of the fatalities in the early period following stroke (42). In the United States, hospital discharge data indicate that in-hospital mortality due to stroke has decreased across the decade 1998 to 2007 (odds ratio (OR) 0.75) (94). Among the common medical complications of stroke, the incidence of pneumonia has not changed; however, the incidence of acute myocardial infarction and urinary tract infection has risen modestly (OR 1.39 and 1.18, respectively). The incidences of deep vein thrombosis and pulmonary embolism have risen more (OR 1.69 and 2.39, respectively).

Cardiovascular events. Stroke is frequently complicated by cardiac events such as myocardial infarction and arrhythmia (60). All patients who are admitted with acute stroke should have an ECG to identify active or prior cardiac ischemia and to assess cardiac rhythm. If available, continuous cardiac telemetry is a useful means of identifying intermittent arrhythmias, such as atrial fibrillation, that may have important diagnostic implications in the stroke patient.

It is important to be aware that ECG changes are seen frequently in acute stroke and may occasionally be the result of central nervous system injury and the corresponding hyperadrenergic state (78). More severe strokes are associated with impaired autonomic cardiovascular control (36). Commonly seen electrocardiographic changes include ST segment depression, QT prolongation, inverted T waves, and prominent U waves (68). Cardiac enzymes should be obtained if ischemia is suspected, but again, they must be carefully interpreted. A consecutive series of 160 acute stroke patients (of which 140 were ischemic) found elevated troponin levels in 10 (6%). Troponin levels rapidly normalized, and chest pain, Q-waves, and focal akinesis were rare, implying that modestly elevated troponin levels in stroke patients are primarily neurogenic in origin (17). Minimally elevated troponin T values, at or below 1 ng/ml, are not infrequent in patients with large anterior distribution infarcts (31). In cases of confirmed myocardial infarction following stroke, decisions about double antiplatelet therapy, anticoagulation, glycoprotein IIb and IIIa inhibitors, and invasive procedures must take into account the extent and acuity of the cerebral infarction to estimate the excess risk of intracranial hemorrhage or other complications associated with these therapies.

Chest radiography is useful to assess heart size and identify pulmonary edema, which should prompt evaluation and management of congestive heart failure. In patients with significant heart failure, efforts to increase cerebral perfusion with intravenous fluids and cessation of antihypertensive agents may have the opposite effect due to worsening of cardiac output. In these cases, optimization of cardiac output is a reasonable primary goal.

Deep vein thrombosis and pulmonary embolism. Venous thromboembolism is a common and potentially devastating complication of stroke. It has been estimated that pulmonary embolism accounts for up to 25% of fatalities following stroke (45). Risk factors for deep vein thrombosis include advanced age, lower extremity paralysis, and atrial fibrillation. Subcutaneously administered heparin or low-molecular-weight heparin have been shown to reduce the risk of deep vein thrombosis (01). One large randomized trial compared enoxaparin (Lovenox, a low-molecular-weight heparin) subcutaneously once daily to unfractionated heparin 5000 U subcutaneously every 12 h for 10 days and reported that enoxaparin reduced the risk of venous thromboembolism by 43% (p=0.0001) (86). There were no significant differences in the risk of any bleeding or symptomatic intracranial and major extracranial hemorrhage. There is also evidence that elastic compression stockings, aspirin, and sequential compression devices can reduce the risk of deep vein thrombosis, and this effect may be additive with anti-thrombotic medication (43; 45). Stroke patients with diminished mobility should receive some form of prophylaxis against deep vein thrombosis or pulmonary embolism. The benefit of mechanical prophylaxis, which is also generally recommended, is more controversial (71). In patients suspected of harboring deep vein thrombosis, lower extremity ultrasound is indicated. Serum D-dimer is probably a less useful alternative, given the high pre-test probability of deep vein thrombosis or pulmonary embolism in stroke patients. Patients with established deep vein thrombosis or pulmonary embolism should generally be treated with anticoagulant therapy. If anticoagulant therapy is contraindicated, consideration should be given to placement of an inferior vena cava filter, though the long-term effects of this intervention are controversial (47).

Infection. Infection is commonplace in the acute stroke patient. Pneumonia and urinary tract infections occur most frequently, each affecting about 10% of acute stroke patients (42). A systematic review finds that infection complicates care in 30% of acute stroke patients (98). Less frequent infectious processes include cellulitis and sepsis. Acutely ill patients, particularly the elderly, may not immediately generate fever and may instead present with unstable vital signs, decreasing oxygenation, or a change in neurologic exam. A high index of suspicion for infection must be maintained and a rapid evaluation of possible sources initiated when appropriate. Close attention should be paid to urinary catheters and intravenous and central lines. Empiric antibiotics are reasonable for patients in whom infection is strongly suspected pending the results of diagnostic tests. However, a randomized trial of 3 days of levofloxacin in acute stroke patients as prophylaxis against infection on admission to the hospital found that the patients who received antibiotics actually did worse than those who did not receive them (18). Aggressive treatment of fever is indicated, as described previously.

Malnutrition and aspiration. Poor nutritional status at hospital admission is associated with worse outcomes following stroke due to increased risk of infection, gastrointestinal bleeding, and bed sores (27). Further, acquired malnutrition following a stroke is also associated with a worse prognosis (21). Swallowing dysfunction is common in acute stroke patients and is associated with a high risk of inadequate nutritional intake and aspiration (54). Swallowing evaluation should be performed in all patients with dysarthria, aphasia, or facial, buccal, or lingual weakness. Inability to swallow safely should precipitate early placement of a nasogastric tube in order to assure gastrointestinal access for nutrition and medications. If swallowing difficulties persist greater than 1 to 2 weeks a percutaneous gastrostomy tube should be considered.

Airway compromise. Patients with brainstem stroke or a decreased level of arousal due to large hemispheric stroke have an increased risk of airway compromise secondary to loss of protective reflexes or oropharyngeal weakness (49). Patients who are unable to protect their airway should undergo endotracheal intubation, with the recognition that those who require intubation have a poor prognosis, with a mortality rate of over 60% (15).

Activity level. Rapid mobilization after the acute period may reduce the risk of pneumonia, deep vein thrombosis, pulmonary embolism, and pressure sores. Shorter time to mobilization and rehabilitation training has been associated with a greater odds of discharge to home (as opposed to a nursing home or death) at 6 weeks (39). As soon as a patient has become neurologically stable, physical therapy should be initiated.

Prevention and management of neurologic complications. Neurologic complications following acute stroke include cerebral edema, recurrent stroke, intracranial hemorrhage, and seizures. Frequent neurologic examinations, followed by repeat neuroimaging if a significant change is noted, may allow rapid identification and treatment of these complications.

Cerebral edema. Patients with large hemispheric or cerebellar infarction are at highest risk of developing clinically significant cerebral edema. Edema typically peaks between 3 and 5 days after stroke onset. Clinical-pathologic correlates include decreased level of consciousness from compression of thalamic and midbrain reticular activating system; enlargement of the ipsilateral pupil from oculomotor nerve compression, ipsilateral lower extremity weakness, or hyperreflexia from subfalcine herniation with compression of the previously unaffected anterior cerebral artery; and ipsilateral upper and lower extremity weakness or hyperreflexia due to Kernohan notch phenomenon (74). Acute hydrocephalus from occlusion of cerebrospinal fluid drainage pathways may be seen with large cerebellar infarctions.

Management of patients with cerebral edema and increased ICP from ischemic stroke is empiric, with limited data to support any particular strategy (01). In general, invasive ICP monitoring does not appear to be of great utility (82). Raising head position to 30 degrees or greater may increase venous drainage and decrease ICP but may decrease cerebral perfusion. Hyperventilation can rapidly reduce ICP, but the effect is short-lived (on the order of hours) and, therefore, of limited utility in the absence of definitive therapy to lower ICP. Osmotic agents such as mannitol, glycerol, and hypertonic saline, which have a longer duration of effect than hyperventilation, may be used to “buy time” until edema begins to subside spontaneously. A common strategy is to give a bolus of mannitol 1.0 g/kg, followed by 0.25 to 0.5 g/kg every 4 to 6 hours for several days or until clinical or radiographic evidence of decreased edema is present. There are no convincing data that these agents improve outcome. Steroids have been shown to be ineffective in ischemic stroke and may increase the rate of infectious complications (63).

In patients with large hemispheric stroke, hemicraniectomy appears to decrease both mortality and morbidity, though the benefit appears to be greater for survival. A pooled analysis of three small randomized controlled trials of hemicraniectomy within 48 hours of stroke found a higher survival rate (78% vs. 29%) and higher percentage of patients with modified Rankin score of 3 or less, indicating less disability (43% vs. 21%) (95). This pooled analysis was likely too small to define differences in subgroups; thus, optimal patient selection and timing of this intervention remains uncertain. However, other studies have suggested that younger patients are more likely to achieve a good outcome, and earlier interventions may be best (83). In patients with large cerebellar infarcts who experience decreasing mental status secondary to brainstem compression and hydrocephalus, there is general agreement that suboccipital craniectomy or ventriculostomy is indicated (56).

Recurrent stroke. Combined analysis of over 40,000 patients in the International Stroke Trial and Chinese Acute Stroke Trial shows that aspirin (160 to 325 mg) given within 48 hours of stroke onset reduces the rate of recurrent ischemic stroke from 2.3% to 1.6% (p< 0.000001) over 2 to 4 weeks. The rate of intracerebral hemorrhage was slightly increased (0.8% vs. 1.0%, p=0.07), but this did not negate the benefit from early aspirin (19). In contrast, unfractionated heparin and low molecular weight heparins have not shown overall benefit in reducing recurrent stroke, partly because of an increased risk of intracranial and other hemorrhagic complications (20). These findings apply to patients in atrial fibrillation at the time of stroke as well (08). In patients who receive thrombolytic therapy, aspirin should be avoided during the first 24 hours (02).

Intracranial hemorrhage. Patients with acute ischemic stroke are at significant risk of intracranial hemorrhage, especially if treated with anticoagulant or thrombolytic therapy. Small asymptomatic hemorrhages may not require any change in therapy. More significant and symptomatic intracranial hemorrhage mandates cessation of anti-thrombotic therapy and reversal of anticoagulation, if applicable. Patients who have received thrombolytic therapy should be treated with cryoprecipitate and platelet transfusion, as per the NINDS rt-PA Stroke Study protocol. A large randomized trial found that surgical evacuation does not improve outcomes in unselected patients with primary intracerebral hemorrhage (58). Nevertheless, a neurosurgical consultation should be obtained and it is reasonable to consider emergency surgical evacuation in patients with large (> 3 cm) cerebellar hemorrhages, or those with large, superficial hemorrhages causing substantial mass effect, with rapidly deteriorating condition.

Seizures. Seizures in the acute ischemic stroke setting have been reported in 1% to 6% of patients (50). Early seizure occurrence has been associated with stroke location (particularly with regard to cortical involvement), size, severity, and hemorrhage (13). Border zone infarction associated with severe carotid artery stenosis has been implicated as a risk factor for early seizure (22). Among 581 young cryptogenic stroke patients, 14 (2.4%) had seizures within 7 days of their event (51). Twenty patients in the overall cohort of 581 patients (3.4%) subsequently developed late seizures (at a mean of 12.9 months), of which only six previously had an early seizure. Four of these six patients had their late seizure while they were already on an anticonvulsant medication. Early seizure was a risk factor for later seizures (hazard ratio=5.1, 95% CI 1.8 to 14.8), a finding that has been reported in another study as well (90). However, the largest collection of ischemic stroke patients published to date (13), with1632 patients, did not find early seizure to be a risk factor for recurrent seizures. It remains controversial as to whether anticonvulsants should be initiated in all patients with early seizure as the potential for seizure prophylaxis should be weighed against factors such as age, disability, and risk of adverse effects (16). Most neurologists find it difficult not to treat patients with seizure in the setting of acute stroke with anticonvulsants; however, continuing therapy for very long in the absence of recurrent seizure seems not to be justified. There is no evidence to support prophylactic use of anticonvulsants in patients with ischemic stroke who have not had seizures.

Organizational strategies: stroke units and transitional care. Specialized units for the care of stroke patients arose in the late 1970s and early 1980s. Such units typically offer continuous cardiovascular and respiratory monitoring and are staffed by a team of healthcare providers—vascular neurologists, neurologic nurses, rehabilitation experts—who have specialized expertise in stroke care. The organizational breadth of these units varies considerably based on geography; in Europe, such units generally encompass the prolonged rehabilitation care of patients, whereas in the United States, they more typically provide care for a shorter time period, with patients transitioning to separate rehabilitation units. These differences complicate interpretation of studies that have been performed to evaluate the effectiveness of stroke units. Nevertheless, randomized trials have consistently demonstrated that stroke units decrease mortality, increase the likelihood of being discharged to home, and improve functional status and quality of life (40). These benefits are durable, persisting up to 10 years after discharge (38). A large meta-analysis of 19 trials with 3249 patients comparing stroke units to general medical wards found that stroke units reduced mortality, death or institutionalization, and death or dependency without a significant increase in length of stay (91). Although many trials have not detected a differential benefit based on stroke subtype, there is some evidence that small vessel infarctions may benefit less compared to other stroke subtypes (25). In one of the studies by Bettger and colleagues, transitional care from hospital to home varied widely among hospitals and there was no single strategy applied universally or provided across hospitals (10). Gesell and colleagues have shown that comprehensive post-acute stroke services were associated with better functional status (30). However, the implementation seems diverse with some challenges, especially reaching out to the patients as well as making sure there is consistent delivery of follow-up visits.

Conclusion. The period immediately following an acute ischemic stroke is a time of significant risk. Meticulous attention to the care of the stroke patient during this time can prevent further neurologic injury and minimize common complications, optimizing the chance of functional recovery. There exists a large gap among eligible patients and the low utilization rates of thrombolysis and mechanical thrombectomy for these patients. Multiple initiatives across the globe are underway to improve systems of care and bridge this gap.