Chapter 5: Acute Ischemic Stroke

Acute ischemic stroke is the most prevalent neurologic emergency in most of the world; one American has a stroke every 40 seconds. In the United States alone there are more than 780,000 strokes per year, with the majority being new events.¹ The number of hospitalizations in the United States continues to increase. The cost associated with the care of stroke patients in 2008 was $65.5 billion; the cost of care per patient is almost double for severe strokes. Stroke is the third leading cause of death in the United States, and among adults it is the leading cause of long-term disability. Stroke is disproportionately a disease of individuals of lower socioeconomic status, African Americans, the elderly, and women in the older age groups. Ischemic stroke accounts for the majority of stroke subtypes in case series from the United States.¹

The majority of stroke survivors have some form of a residual disability, although 50% to 70% will nonetheless regain functional independence.¹ These patients remain at high risk for subsequent morbidity and mortality. A small proportion of acute ischemic stroke patients will be eligible for reperfusion therapy, and an even smaller proportion will actually receive it. In series from various locations in the United States the rate of thrombolysis varies when considering all stroke patients but remains low at 2.0% to 8.5%,² while analysis of data from the Nationwide Inpatient Service reveals this rate to be less than 2%.³ The primary reason for not receiving reperfusion therapy is arrival outside of the appropriate time window.⁴ Those who arrive by the appropriate window still have several reasons within the national guidelines for not being treated, and in some hospital series there are sizable numbers of patients who do not meet any exclusion criteria but are still not treated.⁵ Prevention and treatment of the complications related to stroke remains the cornerstone of treatment for ischemic stroke.

Risk factors for ischemic stroke are somewhat similar to those of ischemic heart disease, with some notable exceptions. Hypertension remains the most important risk factor for ischemic stroke.⁶ Dyslipidemia, while prominent for ischemic heart disease, is less important as a risk factor for ischemic stroke.⁷ Large population-based cohort studies have failed to find a consistent association between dyslipidemia and ischemic stroke, although for the atherosclerotic stroke subtype, this may still be the case.⁸

Atherosclerotic disease of the extracranial or intracranial vessels is an important cause of ischemic stroke, most likely due to artery-to-artery emboli, or in a smaller proportion of cases due to flow failure.⁹ Atrial fibrillation is an important risk factor, particularly for the older population, and frequently leads to more severe strokes and a poorer response to thrombolysis. Atrial fibrillation is more likely to be associated with large infarctions that could lead to malignant cerebral edema, as well as hemorrhagic infarcts/transformation.

Stroke remains a clinical diagnosis, and in the acute setting a history and physical examination remain integral components of evaluation. In the triage area staff can activate the stroke team, if they have not already been activated based on a notification from the EMTs, and start the initial evaluation. This will include checking a fingerstick glucose level, vital signs, and a screening examination such as the Cincinnati Stroke Scale.¹⁰ The latter can be used by nonphysician personnel with excellent reproducibility and sensitivity and involves checking for a facial droop, arm drift, and dysarthria. In patients with suspected stroke, two large-bore intravenous (IV) lines (at least 20 gauge) should be placed, and the following serum tests should be sent to the laboratory immediately: complete blood count, coagulation panel, basic metabolic panel, troponin, and type and hold. Our acute stroke pathway involves completing the history, carrying out an examination, and obtaining neurologic imaging (Figure 5-1).

The history must be focused, and the initial goal should be to establish the exact time of onset. Frequently, patients such as the one above will not be able to provide an exact time of onset, and it is prudent to confirm the time of onset with friends or family. It is important to establish if there were minor symptoms present before the ones that caused the presentation, as well as whether the patient woke up with any deficits. For patients who wake up with a neurologic deficit, the time of onset is assumed to be the moment the patient went to sleep, or otherwise the last time they were seen at their baseline.

A complete neurologic examination is not often necessary in the initial evaluation in the ED; the National Institutes of Health Stroke Scale (NIHSS) is the initial examination of choice. It has excellent interexaminer reliability and sensitivity for most strokes,¹¹ can be performed quickly, and training in its use is available free of charge through the American Heart Association. Depending on the clinical scenario, more detailed examination may be required for suspected nondominant hemisphere injury, subtle aphasia, or cerebellar syndromes with a predominant astasia-abasia picture. Neurologic imaging is obtained before or after the examination, but should be completed within 20 minutes of arrival at the ED.

Given that stroke remains a clinical diagnosis, a noncontrast head computerized tomography (CT) scan is the initial test of choice for most patients.¹² The primary goal of the noncontrast head CT is to rule out intracerebral hemorrhage (ICH). In acute ischemic stroke the head CT will frequently be normal. Early infarct signs can be seen in the head CT and include loss of the insular cortical ribbon, sulcal effacement, and loss of the gray-white junction with blurring of basal ganglia nuclei differentiation. Occasionally, the dense middle cerebral artery sign, which likely represents intravascular dense clot material, is seen (Figure 5-2). Frequently, adjustment of the contrast may be required to visualize the early infarct signs.

In many stroke centers in the country, CT angiography and perfusion are routinely obtained in all acute ischemic stroke patients. CT angiography has the primary goal of identifying patients who have a major arterial occlusion, which may not respond well to intravenous thrombolysis, or which may portend risk of a large hemispheric infarction. CT perfusion is often obtained with CT angiography (Figure 5-3) and helps identify tissue that may be ischemic, but not yet infarcted, and therefore salvageable. CT perfusion allows the measurement of three parameters: cerebral blood volume (CBV), mean transit time (MTT), and cerebral blood flow (CBF). In early cerebral hemodynamic failure, the CBV will increase as distal arterioles vasodilate as a means of maintaining CBF in the face of dropping cerebral perfusion pressure (CPP). When this is insufficient, the CBV may remain increased or begin decreasing to normal or below, and the oxygen extraction fraction will increase; in the next stage, the cerebral metabolic rate of oxygen drops, and soon thereafter CBF will also drop.^13,14 The CBV and CBF maps are particularly useful in detecting potential tissue at risk, particularly when the CBF is low-normal and the CBV is increased. A very low CBV alternatively may indicate already infarcted tissue. The MTT is a useful adjunct, but there may be significant asymmetry in the MTT maps without having tissue at risk, and the optimal threshold for defining perfusion failure remains controversial.

Administering contrast remains a concern for many radiologists and ED physicians without knowing the patient's serum creatinine value. Unless there is a clinical history of renal disease or diabetes, the risk of developing contrast nephropathy is low,¹⁵ and contrast can be administered without knowing the serum creatinine level; if there is still concern, isotonic sodium bicarbonate infusion can be administered. In one study, 3 ampules of sodium bicarbonate were mixed in 1 L of D₅W, and 1 hour before contrast the patient received 3 mL/kg, followed by 1 mL/kg for the 6 hours after the contrast is administered.¹⁶ N-acetylcysteine 600 mg up to three times daily is frequently administered, but its clinical effectiveness may not be as high.

Transcranial Doppler (TCD) is commonly used in acute stroke evaluation to diagnose a major cerebral artery occlusion that may dictate advanced treatment, and may have a therapeutic benefit as well. The presence of a portable TCD machine in many institutions has made this a practical diagnostic tool. Magnetic resonance imaging (MRI) is available acutely in some centers, but in most centers it is not used before a decision to provide treatment in the acute setting. MRI can be useful to diagnose stroke mimics, distinguish TIA from stroke, and evaluate for the presence of a large infarction which may require a different degree of monitoring. MRI, however, is not necessary to make a recommendation on thrombolysis. The diffusion-weighted image (DWI) identifies areas of cytotoxic edema and infarction, although some DWI-positive lesions can be reversible. At the cellular level DWI identifies areas of impaired water molecule diffusion, frequently resulting from energy failure and an inability to maintain membrane gradients, such that brownian motion of water molecules is restricted.¹⁷ Magnetic resonance angiography provides similar information to computed tomographic angiography (CTA), but it may overestimate the degree of stenosis. Magnetic resonance perfusion (Figure 5-4) may be useful in some clinical scenarios to establish a perfusion-diffusion mismatch, but the optimal parameters to define the penumbra remain an area of controversy.

The use of biomarkers in the diagnosis of stroke has received considerable attention in the literature, particularly for diagnosing ischemic stroke, but also for prognosis. Inflammatory markers (C-reactive protein) and measures of glial and endothelial cell injury (matrix metalloproteinase-9, S100B) have been associated with a diagnosis of ischemic stroke, although further data are required to validate their routine clinical use.¹⁸

Intravenous Thrombolysis

Figure 5-5 outlines a suggested treatment paradigm for eligible patients. The only proven therapy to improve clinical outcomes for acute ischemic stroke is IV recombinant tissue plasminogen activator (r-tPA). The pivotal National Institute of Neurological Disease and Stroke (NINDS) r-tPA trial indicated that in patients treated within 3 hours of acute ischemic stroke there were significant clinical benefits.¹⁹ The NINDS r-tPA trial did not demonstrate clinical effectiveness at 24 hours, although there was a trend toward benefit. The significant benefit was gleaned at 3 months where r-tPA was superior to placebo in the primary outcome of a composite of the NIHSS, modified Rankin Scale, Barthel Index, and Glasgow Outcome Scale. The treatment arm was associated with a 13% absolute benefit of attaining minimal to no disability after stroke (39% versus 26%), which translated to a number needed to treat of close to 8. There was no difference in mortality rate. The exclusion criteria for r-tPA need to be reviewed before administration and are summarized in Table 5-1.²⁰ There are several relative contraindications, such as hypodensity in greater than one-third of the middle cerebral artery (MCA) distribution and major deficits, age older than 80 years, and minor or rapidly resolving symptoms. The patient would be eligible for IV r-tPA, but his blood pressure should be lowered to less than 185/110 mm Hg before treatment with either IV pushes of labetalol or the use of an IV nicardipine drip.²⁰

The NINDS r-tPA study differed from other acute stroke treatment trials by narrowing the therapeutic window to 0 to 3 hours from stroke onset, and using a dose of 0.9 mg/kg of r-tPA. Since then, analyses from this and other trials have pointed to a greater benefit to earlier treatment, particularly when r-tPA is started within 90 minutes of stroke onset.^21,22 Thrombolysis with r-tPA is beneficial in all subgroups and stroke subtypes, and its effectiveness has been proven in community cohorts. The typical protocol in the United States includes a door-to-needle time of no more than 60 minutes, with administration of r-tPA 0.9 mg/kg using a 10% bolus given over 1 to 2 minutes, followed by the rest of the infusion over 60 minutes. Consent is not required for IV r-tPA, as it is considered the standard of care, but the patient and family should be made aware that the treatment is being provided.²³ Thrombolysis can be safely and successfully administered via telemedicine, which is an attractive means for increasing treatment in rural and underserved hospitals.²⁴ Delays while waiting for the coagulation profile to return are common, but r-tPA can be safely administered without the coagulation profile.²⁵ After thrombolysis the blood pressure should remain less than 180/105 mm Hg, and any unacceptably high pressure should be promptly treated with easily titratable intravenous antihypertensive medications such as labetalol or nicardipine.

Intravenous treatment with thrombolysis is not effective in many individuals, and has a low likelihood of achieving vessel recanalization, particularly in individuals with an occlusion in one of the major intracranial arteries. Up to 34% of patients treated with r-tPA experience reocclusion after recanalization, with a subsequently high risk of neurologic worsening; these patients, however, still have better outcomes than those who do not recanalize.²⁶ Sonothrombolysis has been one proposed means of improving recanalization. High-frequency sound waves can cause cavitation and the formation of small bubbles that vibrate at a sufficiently high frequency to disrupt the fibrin network in thromboemboli, thereby facilitating entry of a thrombolytic into the clot. In the combined lysis of thrombus in brain ischemia using transcranial ultrasound and systemic tPA (CLOTBUST) phase II study, 126 patients who received IV r-tPA received placebo or a continuous insonation of the MCA via TCD probe.²⁷ The primary endpoint of vessel recanalization was achieved in 46% of the treatment arm versus 18% in placebo arm, with the benefit being maintained at 36 hours; in a secondary outcome there was a trend toward improved functional outcomes, and no differences in safety. Including microparticles to help break up the clot has been an active area of research. In one pilot study, microparticles were able to penetrate MCA thrombi and pass beyond,^28,29 while others have demonstrated improved recanalization rates with the use of microbubbles.³⁰ The type and composition of the microbubbles remain under investigation. Other investigators have attempted to combat the difficulties of vessel reocclusion or to potentiate the effects of r-tPA by adding pharmacologic adjuncts. The combined approach to lysis utilizing eptifibatide and r-tPA in acute ischemic stroke (CLEAR) trial randomized 94 patients to IV r-tPA at a reduced dose (0.3 mg/kg or 0.45 mg/kg) and eptifibatide versus IV r-tPA alone at full dose (0.9 mg/kg), with the primary end point being symptomatic intracranial hemorrhage (sICH) at 36 hours.³¹ There was no statistically significant difference in the safety outcome or clinical effectiveness at 3 months, and until further trials are performed there is no evidence to support this approach. Holding antithrombotics for at least 24 hours after thrombolysis is currently advised.²⁰

The benefit of r-tPA has been robust in routine clinical use, although there is still some resistance toward its use. Relative contraindications for r-tPA use include difficult to control blood pressure, “too good to treat” or rapidly improving syndromes, age older than 80 years, and seizure at onset. Blood pressure that requires aggressive control is not associated with increased rates of hemorrhage in at least one retrospective review,³² while the benefit of aggressive glycemic control after thrombolysis remains under investigation. Several investigators have found that a substantial proportion (close to 20% in most case series) of the “too good to treat” patients end up with poor neurologic outcomes.^33-35 It remains an active area of research to identify which of these patients will worsen and which will not, although a more severe early benefit, primary motor symptoms, or the presence of a major cerebral artery occlusion appear to be notable risk factors for a poor outcome in the “too good to treat” group.^35-37 However, other case series have demonstrated that these patients may not have a poor outcome.³⁸ On the other hand, the safety of thrombolysis has been established for stroke mimics and these mild stroke patients.³⁹ Whether these “too good to treat” or rapidly improving patients should be treated remains a clinical decision on a case-by-case basis, with decision making being driven by the clinical impression of the disability the patient would have if left untreated. There is additional controversy as to whether individuals who are older than 80 years should be treated, although they still gain clinical benefit.^40,41 Several case series have reported an acceptable safety profile in the pediatric population,^42,43 although the clinical benefit in the pediatric population remains unknown.⁴⁴ Intravenous r-tPA remains beneficial even in patients with an initially severe deficit, and it appears to be safe in patients with cervical artery dissection, seizure at onset, and the presence of abnormality on CT; case series have documented it to be safe in patients with a known unruptured atrioventricular malformation (AVM) or aneurysm.⁴⁵ IV r-tPA should not be withheld for an otherwise eligible patient with an incidental finding of an unruptured aneurysm.

Whether IV r-tPA should be withheld from patients with other exclusion criteria is an additional area of controversy. A recent report demonstrated the safety at a single center of administering r-tPA to any patient arriving under 3 hours without a hemorrhage on head CT, while another report documented a high intracerebral hemorrhage rate after thrombolysis in patients taking warfarin with an International Normalized Ratio (INR) less than 1.7.⁴⁶ Regardless, the primary reason for not receiving intravenous thrombolysis remains arrival outside of the appropriate time window.² Interestingly, the bulk of the litigation cases in the United States related to r-tPA are related to failure to administer the medication.⁴⁷

The most dreaded complication from r-tPA infusion is hemorrhage, of which ICH is the most serious. It will typically present with headache, nausea and vomiting, worsening of the neurologic deficit, and in more severe cases an altered level of alertness. In the NINDS tPA trial, the rate of symptomatic ICH (sICH), defined as the presence of hemorrhage on CT of the head with suspicion for hemorrhage or a decline in neurologic status, was present in 6.4% of those receiving r-tPA and 0.6% in those receiving placebo.¹⁹ Of those who had sICH in the r-tPA arm, close to 50% had died at 3 months. Asymptomatic ICH was present in 4.4% overall. Major systemic hemorrhage was rare, while minor extracranial hemorrhage occurred in 23% of patients treated with r-tPA (compared to 3% in the placebo arm).

In the European Cooperative Acute Stroke Studies (ECASS), hemorrhages were classified as follows (Figure 5-6): Hemorrhage infarction type 1 (HI-1), small petechiae along the margins of the infarct; hemorrhage infarction type 2 (HI-2), more confluent petechiae within the infarcted area but without a space-occupying effect; parenchymal hematoma type 1 (PH-1), a hematoma in less than 30% of the infarcted area with some space-occupying effect; and parenchymal hematoma type 2 (PH-2), a hematoma in greater than 30% of the infarcted area with substantial space-occupying effect or as any hemorrhagic lesion outside the infracted area.⁴⁸ PH-2 is the only one associated with clinical decline.⁴⁹ HI-1 and HI-2 are not predictive of a poor outcome, but their presence may indicate successful reperfusion and an association with subsequent clinical benefit from thrombolysis.⁵⁰

The risk factors for development of sICH after thrombolysis have been variable, depending on the study. The most consistent risk factors across several studies include early hypoattenuation on head CT, elevated serum glucose and history of diabetes, hypertension,⁵¹ increased stroke severity and size of stroke on DWI,⁵² especially if there is reperfusion,⁵³ and protocol violations with treatment outside of the time window.^{52, 54-56} It remains controversial whether treatment with antiplatelet agents before thrombolysis,^57,58 pretreatment with statins,^59,60 the presence of microhemorrhages on T2^∗ sequence or gradient echo of MRI,^61,62 leukoaraiosis,⁶³ advanced age,⁴¹ hemostatic factors,^64,65 early disruption of the blood-brain barrier via permeability imaging,⁶⁶ markers of endothelial cell injury (matrix metalloproteinase 9, S100B),¹⁸ and persistent arterial occlusion⁶⁷ convey an additional risk for sICH. The presence of early infarct signs correlated with more severe stroke in the NINDS tPA trial, although they did not mitigate the clinical benefit of treatment or alter the safety outcomes.⁶⁸

Perfusion studies may play an additional role in identifying those patients who are at high risk of sICH. The presence of a large area of reduced CBF as measured by xenon CT correlated well with an increased risk of sICH,⁶⁹ while a very low CBV may be a better predictor than absolute infarct size or relative ischemia.⁷⁰ There are insufficient data at this time to recommend withholding thrombolysis for patients with these characteristics, although they may be ones for whom more intensive monitoring in the first 24 to 48 hours is warranted.

Management of sICH after r-tPA infusion usually starts with discontinuing the infusion of thrombolytics in those patients for whom there is a clinical suspicion of hemorrhage, followed by immediate noncontrast head CT and full coagulation panel including fibrinogen and a complete blood count. By the time sICH is detected on CT scan, most patients have completed their r-tPA infusion. Unfortunately, there is no proven therapy to reverse the effects of r-tPA. However, if active bleeding is found, options include fresh-frozen plasma (FFP) and platelet transfusions or even recombinant factor VII on a case-by-case basis.

Another rare clinical complication of r-tPA infusion is angioedema, which appears to be caused by a similar pathway that has been implicated in angiotensin-converting enzymes (ACEs). It typically occurs 30 to 120 minutes after the infusion of r-tPA in typically 1% to 3% of patients, and curiously enough it tends to occur contralateral to the ischemic lesion.⁷¹ Angioedema needs to be distinguished from tongue hematoma, which has also been reported. Activation of the complement and kinin cascades due to the presence of increased concentrations of plasmin have been implicated. This latter pathway is implicated in an increased risk being present in patients taking ACE inhibitors.⁷² It seems reasonable to assume that patients who develop angioedema may be at an additional risk of the same complication with ACE inhibitors in the future. Management is based on the paradigms from case series, and includes administration of diphenhydramine 50 mg IV and H2-blocker as first-line treatment, followed by 100 mg IV of methylprednisolone or nebulized epinephrine. In the more severe cases, the r-tPA infusion should be stopped in light of a possible loss of the airway and the potential need for endotracheal intubation. The latter may require fiberoptic assistance in severe cases, such as in cases exhibiting stridor and airway compromise. An emergency tracheostomy may be required.

There are additional safety concerns about r-tPA as a potentially neurotoxic substance.⁷³ Animal studies indicate that r-tPA can cross the blood-brain barrier⁷⁴ and, once within the brain parenchyma, can increase ischemic injury via potentiation of N-methyl-d-aspartate (NMDA)–induced cell death and increase NMDA-mediated intracellular calcium levels.⁷⁵

The Alteplase Thrombolysis for Acute Noninterventional Therapy in Ischemic Stroke (ATLANTIS)^76,77 and ECASS I⁷⁸ and II⁷⁹ trials were notable for an attempt to try to expand the window for thrombolysis or alter the dose. In all of these cases, r-tPA was not associated with a clear clinical benefit when the time window extended beyond 3 hours, and appeared to increase the risk of ICH. The European Medicines Evaluation Agency approved r-tPA in 2002 under the condition that a trial of r-tPA be carried out and that a prospective registry include safety outcomes.⁸⁰ This registry tracked patients treated up to 4.5 hours, and there was no difference in complication rates between those who were in the less-than-3-hour window and those in the 3.0- to 4.5-hour window.⁸¹

In 2008, publication of the results from ECASS III revealed a statistically significant clinical benefit in a select group of patients between 3.0 and 4.5 hours from onset.⁸² The trial excluded patients with an NIHSS greater than 25, those older than age 80, those taking oral anticoagulants regardless of prothrombin time, and those with a prior stroke and concomitant diabetes. The trial randomized 821 patients to r-tPA 0.9 mg/kg or placebo, with a primary outcome of mRS (modified Rankin Scale) at 90 days that was achieved in the treatment arm (52.4%) versus placebo (45.2%). Treatment was not associated with increased mortality rate, although there was a higher risk of hemorrhage (2.4% versus 0.3%). A science advisory from the American Heart Association/American Stroke Association recommended that IV r-tPA be administered up to 4.5 hours from stroke onset, but warned against prolonging time windows in the ED for those under 3 hours from stroke onset.⁸¹ A proposed clinical pathway for patients between 3 and 8 hours from stroke onset is presented in Figure 5-7.

In order to increase the proportions of individuals who receive intravenous thrombolysis, several strategies have been tried to date. Education of the public has achieved moderate success in improving arrival times at the ED, while the establishment of acute stroke teams and experience over time has led to reduced hemorrhagic complication rates and protocol violations and improved patient flow in EDs. The ischemic penumbra presented an attractive target beyond 3 hours based on retrospective reviews of clinical protocols of thrombolysis using perfusion-weighted imaging (PWI).^83,84 Moving toward a tissue-based window rather than the absolute time and “clock on the wall” window remains the most promising avenue of research to expand the treatment window.

Trials with other thrombolytics besides r-tPA have suggested a possible role of treatment beyond 4.5 hours in patients with perfusion-diffusion mismatch on MRI. Phase II trials have been carried out with desmoteplase, a chemical derivative from the saliva of vampire bats. Desmoteplase acts by a different mechanism of action, and has a high affinity for fibrin without having an effect on plasminogen or fibrinogen or apparent neurotoxicity.^73,75,85 In the Desmoteplase in Acute Ischemic Stroke trial (DIAS),⁸⁶ patients were randomized to placebo versus desmoteplase treatment if they presented in the 3- to 9-hour time window, had an NIHSS of 4 to 20, and evidence of MRI perfusion-diffusion mismatch. In the trial a non–weight-based dose led to excessive symptomatic hemorrhages (26.7%), with a rate that was significantly lower (2.2%) when it was switched to a weight-based algorithm. The treatment arm was associated with improved reperfusion and subsequent clinical outcomes. In the phase II Dose Escalation of Desmoteplase for Acute Ischemic Stroke (DEDAS)⁸⁷ study that included 37 patients, desmoteplase was associated with a trend toward improved clinical outcomes and a favorable safety profile in patients treated in the 3- to 9-hour time window. The findings were not confirmed in the phase III DIAS-2 trial, where desmoteplase treatment showed no improvement in clinical outcomes, and appeared to be associated with increased mortality rate.⁸⁸ In the Diffusion and Perfusion Imaging Evaluation for Understanding Stroke Evolution (DEFUSE) study, patients within 3 to 6 hours from stroke onset received r-tPA regardless of their baseline MRI profile. Patients with a favorable diffusion-perfusion pattern (small DWI lesion, with a perfusion-diffusion ratio of 2.6) and subsequent reperfusion had a favorable clinical response.^89-91 This study included only 74 patients and did not include a placebo arm.

In the larger Echoplanar Imaging Thrombolytic Evaluation Trial (EPITHET),⁹² 101 patients were randomized to placebo versus r-tPA at 3 to 6 hours from stroke onset, with a primary hypothesis of clinical benefit in patients with a diffusion-perfusion mismatch. The primary outcome was infarct growth between baseline DWI and 90-day T2-weighted imaging for which there was a nonstatistically significant trend to benefit with r-tPA. Secondary outcomes included improved neurologic (NIHSS 0 to 1 or 8 or more point improvement at 90 days) or functional status (mRS 0 to 2 at 90 days) that was associated with reperfusion, which was defined as 90% or more improvement in PWI imaging volume at days 3 to 5. Reperfusion appeared to be more important than recanalization.⁹³ Criticisms of these studies relate to the technique used in measuring the penumbra. There are numerous difficulties with obtaining standardized processing of the perfusion sequences, as well as having them processed in a timely fashion for routine clinical use.

Additionally, the imaging of the penumbra was based on using the magnetic resonance perfusion (MRP) by examining the following parameters: mean time to enhance (mean transit time), negative enhancement integral (cerebral blood volume), and maximum slope of decrease (cerebral blood flow, calculated as CBV/MTT). Hypoperfused tissue on PWI has been primarily defined in EPITHET and DEFUSE as a greater than 2-second delay in the Tmax, analogous to the MTT, with reperfusion being defined as a greater than 90% improvement in this parameter. On the other hand, final infarct volume appears to correlate better with a Tmax of 6 to 8 seconds or more.⁹⁴ The other image maps are usually processed with a significant delay and are not readily available. In contrast, further understanding of the penumbra has led to the understanding that some portions of the penumbra will remain viable regardless of whether reperfusion occurs, while other will not, despite treatment.⁸³ The role of collateral flow, and imaging it, remains incompletely understood, as does the optimal parameters to define the penumbra. The correlations with quantitative CBF from other modalities also remain imperfect. It remains important to comprehend that recanalization, which thrombolytics may do, is not synonymous with reperfusion, although they tend to be correlated. The former is based on angiographic improvement in the appearance of the vessel, while the latter refers to an improvement in PWI parameters. A recent meta-analysis showed that reperfusion and recanalization rates improved using mismatch-guided thrombolysis and that recanalization improves outcomes; however, thrombolysis did not improve outcomes.⁹⁵ Currently, it remains controversial whether perfusion-diffusion imaging should be used in routine clinical care.^96,97

Pharmacologic Treatment

Table 5-2 outlines the results of notable pharmacologic and device trials in acute ischemic stroke. Glycoprotein IIb/IIIa (GPIIb-IIIa) inhibitors have been an attractive target for pharmacologic treatment beyond thrombolysis. In a phase II dose-escalation study enrolling 74 patients, investigators found that the administration of abciximab was associated with a trend toward improved disability, with no significant increase in the rate of sICH.⁹⁸ The Abciximab in Emergency Treatment of Stroke Trial (AbESST), a phase II study using 400 patients, found similar results in terms of safety and a nonsignificant modest trend toward improvement in mRS at 3 months.⁹⁹ The phase III AbESST II study planned to enroll 1800 patients, with randomization to abciximab or placebo within 5 hours of stroke onset in those who did not receive thrombolysis. The phase III study failed to replicate these findings, as the study was stopped prematurely after 808 patients were enrolled by the data safety monitoring board. There was a significant increase in the rate of sICH in the treatment arm (5.5% versus 0.5%), with no difference in the prespecified primary outcome.¹⁰⁰ GPIIb-IIIa inhibitors have also been examined in combination with thrombolysis with the intent to improve recanalization rates based on experience with acute myocardial infarction.

Antithrombotic agents remain the most commonly used pharmacologic agent for acute stroke prevention. In the Chinese Acute Stroke Trial,¹⁰¹ aspirin showed a modest absolute benefit over placebo in mortality and recurrent stroke risk. In the International Stroke Trial (IST),¹⁰² aspirin again had a mild benefit over placebo in preventing stroke and improving mortality rate, while the heparin arm showed a trend toward a mild benefit in the 12,500 arm versus the 5000-unit arm that was offset by the hemorrhage risk. Danaparoid, a heparinoid, was found to be ineffective in grouping the acute setting for noncardioembolic stroke,¹⁰³ while low-molecular-weight heparin showed a similar lack of effectiveness in acute cardioembolic stroke.¹⁰⁴ Pilot data from the LOAD (Loading of Aspirin and Clopidogrel in acute ischemic stroke and transient ischemic attack) trial indicated a favorable safety profile in the acute setting for the combination of aspirin and clopidogrel, with a trend of reduced clinical worsening.¹⁰⁵ In the larger FASTER (Fast Assessment of Stroke and Transient ischemic attack to prevent Early Recurrence) trials, there was a reduction in clinical worsening with a combination of the two antiplatelet agents, which offset the slight increase in hemorrhage.¹⁰⁶ Currently, there are no phase III clinical trials proving the effectiveness of aspirin and clopidogrel in combination after acute ischemic stroke, although caution is warranted given the lack of clinical effectiveness in secondary prevention studies.¹⁰⁷ Heparin may be considered in the acute setting in the post–cardiac surgery stroke population, in selected cases of carotid dissection, and in patients with a mechanical heart valve.

Intra-arterial Treatment

Given the poor outcomes associated with not recanalizing the middle cerebral artery in particular, endovascular treatment has emerged as another option for acute stroke treatment. In the PROACT (Prolyse in Acute Cerebral Thromboembolism) study, 40 patients were randomized to intra-arterial (IA) infusion of prourokinase versus placebo. Treatment with prourokinase was associated with a statistically significant increase in the rate of recanalization, although there was a concomitant increase in the rate of sICH in the medication arm (15.4% versus 7.1%).¹⁰⁸ Much of the sICH could be attributed to the use of IV heparin.

Publication of the PROACT II study in 1999 significantly altered the paradigm for acute stroke care, and in specialized stroke centers throughout the country, treatment up to 6 hours from stroke onset became possible, if not part of routine clinical care. In PROACT II,¹⁰⁹ 180 patients were randomized to receive 9 mg of IA prourokinase plus heparin versus heparin alone. The main outcome was an mRS of 0 to 2, which was achieved in 40% of the treated patients versus 25% in the control group, with a 66% versus 18% success rate at recanalization with IA treatment. There was a statistically significant increase in the rate of sICH (10% versus 2%), although the investigators concluded that the benefits of the treatment outweighed the risks. The Middle Cerebral Artery Embolism Local Fibrinolytic Intervention Trial confirmed improved functional (mRS 0 to 1) and neurologic (NIHSS 0 to 1) outcomes at 90 days with intra-arterial urokinase, although the study's primary outcome of mRS 0 to 2 was not reached.¹¹⁰ In PROACT II age, NIHSS, and initial head CT were all associated with a greater likelihood of clinical benefit from intra-arterial treatment, while experience at the University of Houston has identified age, NIHSS, and admission glucose as predictive of a poor outcome independently of recanalization.¹¹¹

The Interventional Management of Stroke (IMS) trials further tested the hypothesis of combining intravenous and intra-arterial thrombolysis for patients who had major intracranial artery occlusions under the premise that these patients are unlikely to recanalize with intravenous treatment alone. The phase I Emergency Management of Stroke (EMS) bridging trial showed that this treatment paradigm was feasible and appeared to be safe.¹¹² In IMS I, 80 patients were treated with 0.6 mg/kg of IV r-tPA followed by intra-arterial r-tPA and were compared to historical controls.¹¹³ The rates of sICH were similar to the treatment arm of the NINDS r-tPA trial (6.3%).

IMS II assigned 81 patients who received IV r-tPA (0.6 mg/kg) within 3 hours to intra-arterial r-tPA up to 22 mg or when recanalization was achieved, as well as the use of an intravascular sonothrombolysis device. Twenty-six patients were treated with IV r-tPA only, while the rest were treated with a combination of intra-arterial sonothrombolysis or microcatheter injections. Despite a nonsignificant trend toward a higher rate of sICH in the treatment arm compared to the NINDS tPA trial (9.9% versus 6.4%), there was a higher likelihood of achieving global outcomes in the endovascular treated group compared to the historical controls. There was an additional trend toward improvement in some of the outcomes in the endovascular treatment arm compared to the NINDS r-tPA–treated patients.¹¹⁴

IMS III is currently enrolling patients using IV r-tPA at a reduced dose (0.6 mg/kg) IV followed by the rest of the infusion in an IA fashion. A criticism of IMS III has been that patients are not receiving the optimal dose of IV r-tPA up front, although in one small clinical trial, patients also had a benefit from treatment with 0.6 mg/kg of r-tPA.¹¹⁵ The optimal dose of IV r-tPA to be given before IA treatment remains controversial, although in at least one case series giving full-dose 0.9 mg/kg of IV r-tPA was safe and associated with successful recanalization.¹¹⁶ The results are eagerly anticipated as there are reports that the combination of IV and IA r-tPA may increase the risk of sICH.¹¹⁷ The risk factors for sICH remain the same for IV and IA r-tPA in many observational studies, with elevated serum glucose remaining a particularly important factor¹¹⁸ and the number of microinjections during IA treatment being an additional factor.¹¹⁹ Intra-arterial thrombolysis up to 6 hours from onset has class I, level B evidence for selected patients with MCA occlusion and class IIa, level C evidence for treatment in patients who cannot receive thrombolysis.²⁰

In the meantime, industry has spearheaded the development of several devices for the endovascular treatment of acute ischemic stroke. These devices have used a combination of clot extraction, clot suctioning, and ultrasound-assisted thrombolysis. There are two currently approved devices by the US Food and Drug Administration for the mechanical clot extraction in an intracerebral artery—the Merci Retrieval System (Concentric Medical Inc, Mountain View, CA) and the Penumbra System (Penumbra, Inc, Alameda, CA). It is important to note that neither one of these devices has been approved to date to benefit patients but rather to recanalize the vessel. In the MERCI and Multi-MERCI studies, a corkscrew-type device was tested up to 8 hours from stroke onset in a nonrandomized fashion in 151 patients. Recanalization was achieved in up to 69.5% of patients, which is higher than in historical controls, with a comparable sICH rate (9.8%). Favorable clinical outcomes, defined as mRS of 0 to 2 at 90 days, were achieved more often in the patients who had successful recanalization.^120,121

The Penumbra device uses continuous suction locally and was tested in 125 patients with an NIHSS greater than 8 and occlusion of a large intracerebral artery in a single-arm fashion. Recanalization was achieved in approximately 83% of participants, which the investigators mentioned was higher than with the Merci device, but there was a 12.8% procedural complication rate of which 2.4% was considered serious. The proportion of patients who had an asymptomatic ICH was 28%, while 11.2% had sICH; 90-day mortality was 32.8%, and 25% of individuals attained an mRS of 0 to 2.¹²²

The window for treatment has been expanded by clinicians for patients who are neurologically devastated from basilar artery thrombosis up to 24 hours with the understanding that the patient's outcome is uniformally dismal, while the risk of hemorrhagic transformation is smaller in the brainstem. One recent research reported superior recanalization success and improved outcomes in patients with acute basilar artery occlusion who were treated with IA thrombolysis in combination with IV abciximab versus IA thrombolysis in a nonprospective manner, and these results cannot be generalized to non–basilar artery involvement.¹²³ A large international registry of patients with acute basilar artery occlusion that examined outcomes after IV r-tPA, antithrombotic treatment alone, and endovascular treatment has provided additional guidance.¹²⁴ In patients with a mild-moderate deficit, IV r-tPA appeared to be superior, while in a severe deficit, IV r-tPA in combination with IA treatment was a promising treatment modality. The investigators could not conclude on the optimal timing for when not to treat patients with IA modalities, although none of the patients treated after 9 hours from stroke onset had a good outcome.

Stenting for acute ischemic stroke, even in the setting of nonatherosclerotic strokes, has also received recent attention. In a 2009 published pilot trial, stenting performed less than 8 hours from stroke onset had an acceptable safety profile compared to other trials, with comparable recanalization rates to other modalities.¹²⁵ Further trials are required with this technology or other emerging modalities such as a retrievable stent.¹²⁶

In patients with internal carotid artery, M1 or M2 segments of the middle cerebral artery, or vertebrobasilar artery occlusions, we will routinely perform catheter angiography after thrombolysis for endovascular treatment, and will consider it for patients up to 8 hours from stroke onset if there is no evidence of large cerebral infarction (more than one-third of the MCA distribution) (see Figure 5-7). In the case of intra-arterial treatment, informed consent is part of our routine clinical protocol, and we will explain thoroughly to the family the risks and benefits, as well as the quality of evidence available.

Nonthrombolytic Acute Stroke Treatment

Statins have shown a clear benefit in secondary stroke prevention, an effect that appears to be independent of the serum lipid panel. The cholesterol-independent effects of statins have been an attractive target in the acute stroke setting. Statins have the potential for being neuroprotectants independently of their effect on modification of dyslipidemia. Animal models have indicated that statins are associated with a reduction in infarct size, a result that was mediated in part by endothelial nitric oxide synthetase expression. Statins also appear to reduce inflammation and may increase angiogenesis, synaptogenesis, and neurogenesis. Lastly statins reduce membrane cholesterol synthesis in neurons, thereby rendering them more resistant to glutamate n-methyl-d-aspartate receptor–mediated excitotoxic effects, one of the principal mechanisms of neuronal death in ischemic brain.¹²⁷ Patients who were taking a statin in the Northern Manhattan Study at the time of their stroke were more likely to have a favorable outcome,¹²⁷ while withdrawal of a statin in the acute stroke setting is associated with worsening outcomes.¹²⁸ There are currently no phase III clinical trials that have enrolled patients in the acute stroke setting. Phase II studies of high doses of statins are currently ongoing.¹²⁹

The American Heart Association (AHA) guidelines recommend that in patients not receiving thrombolysis the blood pressure should be allowed to remain as elevated as 220/120 mm Hg unless there is another medical reason not to do this. The rationale remains that autohypertension may maintain CBF in the face of impaired autoregulation after stroke, particularly in the penumbra. There is the associated belief that lowering the blood pressure could expand the size of the present infarct. The specific clinical trials that have demonstrated this are scant. Investigators have begun to examine this finding in rigorous clinical trials, though in the recently completed Scandinavian Candesartan Acute Stroke Study, treatment with candesartan was associated with a trend toward worse outcomes, primarily because of worsening neurologic status.¹³¹ In phase II trials that enrolled both ICH and ischemic stroke patients, lowering of the blood pressure was not associated with worsening clinical deterioration.¹³⁰

Salvaging the penumbra by improving blood flow to the region has been the rationale for many of the acute stroke trials. An alternative approach has been to reverse the biochemical pathways in the ischemic penumbra so as to prevent the progression to infarction, and thereby salvage more tissue than just with thrombolysis. Many of the agents have been successful in animal models. In specific G_M1 calcium channel antagonists, anti–intercellular adhesion molecule 1 antibodies, G_M1 ganglioside, γ-aminobutyric acid agonists, and sodium channel antagonists have all been tried unsuccessfully.¹³¹ Preventing glutamate excitotoxicity at the level of the NMDA receptor has been a particularly attractive target, as has targeting free radicals. Gavistenel, a glycine antagonist of the NMDA receptor that reduces glutamate receptor depolarization, was not associated with improvement in clinical outcomes in two separate trials.^132,133 In the SAINT (Stroke Acute Ischemic NXY-059 Trial) I trial, patients were randomized to NXY-059 (a free radical scavenger) versus placebo, and there appeared to be a trend toward improvement in mRS at 3 months.¹³⁴ These findings were, however, not replicated in the SAINT II trial.¹³⁵ The reason for the failures of all these clinical trials points in part to the imperfect animal models for ischemic stroke.

Carotid revascularization remains one of the most effective secondary prevention strategies for ischemic stroke in the setting of atherosclerotic stenosis of greater than 70%,¹³⁶ and in many centers carotid endarterectomy (CEA) remains the procedure of choice. Recent controversies have emerged regarding the timing of surgical intervention, and whether endovascular modalities are superior. The optimal modality for revascularization appears to be dependant on patient characteristics. In the recently completed Carotid Revascularization Endarterectomy versus Stenting Trial (CREST) investigators found that carotid angioplasty-stenting (CAS) had a similar rate to CEA for the composite end-point of stroke, myocardial infarction, or death.¹³⁷ Investigators found that CAS was associated with a higher risk of perioperative stroke, while CEA was associated with a higher risk of perioperative MI. Pooled data from United States and European trials support that the greatest clinical benefit is derived from revascularization performed within the first 2 weeks of ischemic stroke,¹³⁸ and recent guidelines support earlier treatment in patients with a minor stroke or transient ischemic attack (TIA).⁶ In patients with a larger area of cerebral infarction, it may be reasonable to wait longer to minimize the risk of hemorrhagic conversion and allow the patient to recover. On the other hand, whether patients with symptomatic intracranial atherosclerotic stenosis should receive endovascular treatment remains in doubt given excessive stroke risk in the endovascular arm of the SAMMPRIS trial.¹⁴¹

The main life-threatening condition for this patient is malignant cerebral edema with subsequent herniation. Complete MCA infarctions can be found in up to 10% of all acute stroke patients, while malignant MCA comprises those patients who have a subsequent decline in their level of alertness and ventilatory capacity; this will typically occur in the first 24 to 48 hours and can peak at days 3 to 5 after the initial injury.¹³⁹ Treatment options for preventing malignant cerebral edema include at this point: (1) Monitoring his neurologic examination every 2 to 4 hours in the NeuroICU or stroke unit to assess for signs of elevated intracranial pressure (ICP); (2) starting hypertonic saline (23.4% boluses recommended as a rescue therapy over continuous infusion of 3% to achieve a certain sodium target) or mannitol boluses (1.0 to1.5 g/kg); and (3) performing an early hemicraniectomy within 72 hours electively. A proposed clinical pathway for management of these patients is outlined in Figure 5-8.

Maintaining normoglycemia has been advocated by some to prevent malignant cerebral edema, but its effectiveness has yet to be proven. Prophylactic normothermia, or even hypothermia to 35.5°C, has also been carried out with the belief that this could spare the need for hemicraniectomy, although the clinical effectiveness remains unproven for prevention of edema. The potential for harm is substantial from shivering, infection, or coagulopathy when used empirically. Treatment with hypertonic saline is unlikely to be effective in preventing malignant cerebral edema in this case and holds the potential for harm.

In addition, although treatment with hypertonic saline may appear to be effective, the patient may become dependent on the treatment and it cannot be withdrawn without precipitating herniation. In addition, continued and prophylactic hypertonic saline may be associated with adverse medical outcomes, and specifically volume overload, lung injury, and hypernatremia, although not all groups have observed these adverse outcomes.¹⁴⁰ Hemicraniectomy is a viable option for this patient; however, the hemicraniectomy trials have required the presence of impaired alertness (most notably a score of at least 1 in the 1a item on the NIHSS) and an NIHSS greater than 15.¹⁴¹ In the trials, treatment needed to be carried out within 45 hours from stroke onset, although there is heterogeneity in this criterion. The most prudent course of action in this case may be watchful waiting for at least 5 days, during which the maximum amount of cerebral edema is expected.

At this time, it should be assumed that the patient has an elevated ICP and midline shift, and repeat a head CT to evaluate midline shift or the presence of hemorrhagic transformation. It has remained difficult to establish which patients with a complete MCA infarction will progress to have a malignant course (Figure 5-9), but it appears that those with first stroke, female gender, young age, heart failure, greater than 50% involvement of the MCA region, carotid occlusion and an abnormal ipsilateral circle of Willis, insufficient leptomeningeal vessels, involvement of the anterior choroidal artery region, and serum biomarkers correlating with infarct volume (such as S100B greater than 1.03 μg/L) are at higher risk.^145-148 The most important predictor may be the overall size of the infarct, and identification of the other factors should not be used to select candidates for hemicraniectomy before there is transtentorial herniation.¹⁴⁶

The treatment options remain the same before the patient becomes somnolent. In this case, the clinical trial data clearly support the use of hemicraniectomy. A pooled analysis of three separate European clinical trials required the previously noted inclusion criteria, as well as greater than 50% involvement of the MCA distribution. In the trials, patients aged 18 to 60 years were randomized to early hemicraniectomy versus maximal medical management and followed for 1 year for the main clinical outcomes of a favorable mRS (0 to 4), as well as mRS of 3 or less and death. The surgical arm had an overwhelming benefit in all of the clinical outcomes. The number needed to treat in these studies was two, and there was a benefit regardless of which side the infarct was on.¹⁴⁴ The trial did not require an absolute cutoff of midline shift, and using an absolute cutoff to trigger hemicraniectomy is potentially dangerous. Predicting the outcome of surgical decompression remains difficult, and there have been no clear associations between outcome and size of the infarct, although in one case series age older than 50 years was a poor prognostic sign.¹⁵⁰ The trials leave many open questions, including the timing of when to operate and whether their outcomes are clinically meaningful to patients and their families. Depending on the patient's age, atrophy, and involvement of the medial temporal lobe, surprisingly small amounts of midline shift can be associated with injury to the thalamus and midbrain.

Medical treatment has been advocated by many clinicians, but the clinical evidence is not particularly strong, and clinical trials and case series demonstrate no change in the high mortality rate in medically treated patients.¹⁴² Treatment strategies in lieu of hemicraniectomy that have been suggested include osmotherapy (with mannitol, glycerol, or hypertonic saline), steroids, barbiturate coma, hyperventilation, and head elevation.^151-153 All of these therapies remain unproven (and transient in effect) in clinical trials, hold the potential for harm,¹³⁹ and should only be used as temporizing measures before surgery. Animal models and early studies indicate that hypothermia remains a promising medical treatment for patients with malignant MCA infarction, although it has yet to be tested in a blinded prospective clinical trial.¹³⁹

The most effective treatment in this condition is emergent surgical decompression, with all other medical interventions being aimed at stabilizing the patient in preparation for the operating room. Rapid intubation is advised, both to prevent pneumonia as well as to help manage ICP. Hyperventilation can be effective at reducing ICP by reducing CBV via vasoconstriction of distal arterioles, although in the long-term this same process could precipitate cerebral injury. Hyperventilation is carried out with the help of an end-tidal Co₂ monitor, with a goal of 30 mm Hg, and for no longer than 30 minutes. After intubation the head of the bed should be maintained at 30 degrees, and sedation should be used to minimize agitation that might increase ICP; use of propofol may be helpful owing to its short half-life, although in select cases dexmedetomidine or midazolam may be considered. If surgery will not be carried out, sedation should be maintained, but at this point an ICP monitor is advisable. Discussion of the type of monitor is beyond the scope of this chapter.

Normothermia is an important component of medical treatment in preparation for surgery, although hypothermia may be necessary for intractable cases of high ICP; prevention of shivering in that case may be important to reduce ICP as well. Strategies to treat shivering will be covered in more detail in other chapters. For rapid lowering of ICP, osmotherapy is carried out at the same time as all of the above interventions. It is advisable to use either (or combination of) a 30-mL bolus of IV 23.4% sodium chloride given through a central line over 5 minutes or 1.0 to 1.5 g/kg of 10% mannitol solution, depending on the volume status. Hypertonic saline in the acute setting may be safer and perhaps more effective based on the trauma literature,¹⁵¹ and has been associated with improvements in multimodality intracranial monitoring parameters^152,153 as well as reversal of transtentorial herniation.¹⁵⁴ Hypertonic saline may be preferable because of a more rapid infusion time; how quickly each agent can be obtained and whether the patient has a central line frequently are the deciding factors in choosing one over the other.

The management of space-occupying infarcts in the posterior fossa is not as clear as in those with middle cerebral artery infarcts. Case series have pointed to a similar risk of cerebellar tonsilar herniation for infarcts of greater than 3 cm in size or those that involve the entire posterior-inferior cerebellar artery distribution, as in cerebellar hemorrhage. However, whether these patients should undergo prophylactic suboccipital decompression is unknown, but the differences in the herniation syndromes between cerebellar and cerebral edema should be considered.

The patient was treated empirically in the ED with maneuvers to improve the blood pressure given the presence of an ICA occlusion. Currently, there are little clinical trial data to support such maneuvers, and many of the published case series are hindered by the lack of blinded neurologic assessments before, during, and after treatment with pressors. Early hypotension in some clinical trials and case series has been associated with poor neurologic outcomes and stroke-related mortality,¹⁵⁵ while the AHA guidelines recommend allowing blood pressure up to 220/120 mm Hg in the acute setting unless there is evidence of end-organ damage.²⁰ In the latter guidelines, pressor-induced hypertension for acute stroke was not recommended for most patients, while reviews indicate there is insufficient evidence for the routine induction of hypertension.¹⁵⁶ However, the cause and effect of hypotension after stroke with poor outcomes has not been established; for example, hypotension could be a manifestation of poor cardiac function, which would predispose patients to adverse outcomes. On the other hand, animal and human studies have established the presence of an ischemic penumbra, where cerebral autoregulation is also impaired and where CBF and cerebral metabolic rate of oxygen can be improved with augmentation of the blood pressure.^155,157 In addition, the recently completed angiotensin receptor blocker candesartan for treatment of acute stroke trial showed a trend toward poor neurologic outcomes in patients treated early with the study medication.¹⁵⁸

Case series in the interim have shown improvements in NIHSS and other clinical measures by treatments with pressors.¹⁵⁹ Few trials of empirically raising the blood pressure in acute stroke have been done. It is important, however, to note that potential dangers do exist with raising the blood pressure, including inducing myocardial ischemia, potentiating pulmonary edema, or worsening neurologic outcome. The latter was explored in the DCLHb study, where ischemic stroke patients were treated in a placebo-controlled safety study with a hemoglobin-based oxygen-carrying solution (DCLHb) that induces brisk increases in blood pressure. In this study, participants treated with the study medication had greater odds of unfavorable outcomes (mRS 3 to 6) as well as medical complications and mortality.¹⁶⁰ This study has guided the latest Cochrane Database System Review to recommend against routine induction of hypertension,¹⁶¹ although it is difficult to gauge if it was the actual agent used versus induction of blood pressure. It may be reasonable to consider inducing hypertension with vasoactive agents in patients who have evidence of a large area of perfusion deficit on imaging and no evidence of myocardial ischemia. In that instance, systematic examinations at preestablished time points with the NIHSS are important, especially those done by blinded adjudicators; these patients should be monitored carefully for cardiac and pulmonary complications.

Other promising areas of clinical care to enhance cerebral blood flow include transient partial aortic occlusion by the use of the NeuroFlo (CoAxia, Maple Grove, MN) device, which is occasionally used for refractory vasospasms after subarachnoid hemorrhage, although this device has yet to be formally tested in large clinical trials.¹⁶²

Neurologic Complications

Patients with large cerebral infarctions may be at risk for other complications beyond herniation. Hemorrhagic conversion after ischemic stroke in patients who do not receive thrombolysis is most strongly associated with a cardioembolic etiology. Hemorrhagic transformation can peak at up to 2 weeks after ischemic stroke. Seizures and status epilepticus in the acute setting are unusual presenting symptoms or complications in ischemic stroke, but any patient with a more severe neurologic examination than the neuroimaging would suggest should be monitored. Antiepileptic medications are not routinely used in ischemic stroke patients, even in those with space-occupying infarcts, particularly due to the concern of the neurotoxic properties of some agents.

Infections

Infectious complications are some of the most likely outcomes in stroke patients and account for a significant proportion of the early morbidity in the care of stroke patients. Pneumonia, usually due to aspiration, is present in as many as 22% of stroke admissions at some point in their hospitalization¹⁶³ and is more likely in patients with dysarthria/dysphasia, age older than 65 years, cognitive impairment, and a failed bedside water swallow evaluation.¹⁶⁴ Urinary tract infections (UTIs) are also very common in stroke patients and present in up to 24% of patients during their inpatient admission.¹⁶³ Bladder dysfunction is very common in the acute period, although in our experience UTIs are more likely to occur as a result of inappropriately prolonged periods of indwelling urinary catheters. Both pneumonia and UTI have been associated with poorer functional outcomes in several studies.^165,166 Some clinicians have advocated prophylactic antibiotics early in the clinical course of patients with large ischemic strokes,^167,168 although this has not been proven in other clinical trials.¹⁶⁹ Admission to an organized inpatient stroke unit is one of the few interventions that reduces in-hospital mortality rate after stroke, and the mechanism appears to be through reduction of infectious complications.¹⁷⁰ Although less frequently documented in clinical studies, we have also observed occasionally life-threatening infections from peripheral IV lines. Our practice is to revisit on a daily basis whether patients require indwelling catheters and IV lines.

Cardiovascular Events

The incidence of ST-elevation myocardial infarction during the hospital stay is rare, but elevated troponins are observed in up to 18% of stroke patients.^171,172 Some myocardial infarctions will manifest as a non–ST-segment elevation event, and may not have associated chest pain, and as such it is often assumed to be secondary to neurocardiogenic origins. Troponin elevations from cardiac stunning are less likely in ischemic stroke than they are in hemorrhagic strokes, and they are most likely to occur in individuals with large hemispheric infarcts. It is still reasonable to assume based on epidemiologic data that patients with ischemic stroke also have atherosclerotic coronary artery disease. Treatment of myocardial infarction after ischemic stroke remains difficult when there are competing interests between the brain and the heart in relation to blood pressure management and antithrombotics. Unless there is an ST-elevation myocardial infarction, we avoid cardiac catheterization because of the possible need for multiple antithrombotic agents; for this same reason we also avoid IV heparin unless there is an imminent need for a cardiac catheterization. Decompensated congestive heart failure is an additional cardiac problem we have observed in our stroke patients, and in most cases it appears to be due to withholding diuretics to induce autohypertension. In general, it is reasonable to continue patients' diuretics and β blockers to prevent pulmonary edema or rebound tachycardia.

Pulmonary Complications

Pulmonary embolus (PE) remains one of the most preventable medical complications of any stroke subtype. The incidence of deep vein thrombosis (DVT) and pulmonary embolus is less than 3% in all stroke admissions.¹⁶³ Treatment with prophylactic heparinoids as early as possible is clearly warranted in all stroke patients unless there is a clear medical contraindication; HI-1 and HI-2 hemorrhagic transformation patients should probably not have their DVT prophylaxis withheld. It appears that the use of compression stockings alone is not sufficient after stroke to prevent DVT.¹⁷³

Respiratory failure after stroke is not only associated with malignant MCA infarcts, but can also be present in other stroke subtypes and independent of cardiopulmonary injury. Independent risk factors for mortality in respiratory failure after ischemic stroke include age older than 60 years and a Glasgow Coma Scale (GCS) score less than 10, while respiratory failure is associated with a low probability of survival (33%) 2 years after stroke.¹⁷⁴ In a sample of patients with stroke and mechanical ventilation in Northern Manhattan Hospital mortality rate, was as high as 50% in ischemic stroke patients and higher in the hemorrhagic stroke subtypes.¹⁷⁵ It is unclear to what degree this is influenced by the decision to withdraw care in these patients by the family, and a cause and effect pathway is difficult to establish. As such, intubation should not be withheld for procedures, particularly when intubation is not felt to be permanently required.¹⁷⁶

Death

In-hospital mortality rate after ischemic stroke is low overall, with one large series from Germany indicating a rate close to 5% and a higher rate being reported in other case series. The causes of death within 30 days after ischemic stroke is heterogeneous, but medical complications remain at the forefront, with pneumonia being a principal cause of death following stroke.^34,165,177 Case series have also described atrial fibrillation as a risk factor for in-hospital mortality, although it is not clear if this is related to the larger infarcts observed in this population or due to underlying cardiac disease.^181-183 Out of all the complications after stroke, the most likely one to be associated with mortality is elevated ICP and cerebral perfusion failure.¹⁷⁷