Key points

Introduction

The treatment of acute ischemic stroke (AIS) shares similarities with other vascular emergencies, in that reperfusion of ischemic tissue, halt in propagation of infarction, and prevention of recurrence are the 3 primary early goals of care. Even more than myocardial and other tissue, however, brain parenchyma is exquisitely sensitive to short periods of oligemia and hypoperfusion. In fact, radiographically proven acute cerebral infarction has been reported in patients with as little as 10 seconds of symptoms. The term “time is brain” has been popularized to emphasize the rapidity by which neurons are irretrievably lost during an ischemic stroke. Although dependent on several factors, including degree of ischemic preconditioning, site of occlusion, perfusion of collateral vessels, blood pressure, and glucose and oxygen delivery, on average 1.9 million neurons are destroyed with each passing minute that a stroke evolves. When translated into patient lifetime benefits from expeditious thrombolysis, each minute saved in stroke onset to treatment led to an average of 1.8 days of additional healthy life (95% prediction interval, 0.9–2.7).

Although stroke recently declined from the third to the fifth most common cause of death in the United States, the annual incidence and overall prevalence continue to increase and it remains a leading cause of long-term disability. Since the available US Food and Drug Administration (FDA)–approved treatment options are time dependent, improving stroke care in the early moments may have more of a public health impact than any other phase of care. Timely and efficient stroke treatment should be a priority for emergency department (ED) and prehospital providers. This article discusses the currently available and emerging treatment options in AIS focusing on the preservation of salvageable brain tissue, minimizing complications, and secondary prevention.

Introduction

The treatment of acute ischemic stroke (AIS) shares similarities with other vascular emergencies, in that reperfusion of ischemic tissue, halt in propagation of infarction, and prevention of recurrence are the 3 primary early goals of care. Even more than myocardial and other tissue, however, brain parenchyma is exquisitely sensitive to short periods of oligemia and hypoperfusion. In fact, radiographically proven acute cerebral infarction has been reported in patients with as little as 10 seconds of symptoms. The term “time is brain” has been popularized to emphasize the rapidity by which neurons are irretrievably lost during an ischemic stroke. Although dependent on several factors, including degree of ischemic preconditioning, site of occlusion, perfusion of collateral vessels, blood pressure, and glucose and oxygen delivery, on average 1.9 million neurons are destroyed with each passing minute that a stroke evolves. When translated into patient lifetime benefits from expeditious thrombolysis, each minute saved in stroke onset to treatment led to an average of 1.8 days of additional healthy life (95% prediction interval, 0.9–2.7).

Although stroke recently declined from the third to the fifth most common cause of death in the United States, the annual incidence and overall prevalence continue to increase and it remains a leading cause of long-term disability. Since the available US Food and Drug Administration (FDA)–approved treatment options are time dependent, improving stroke care in the early moments may have more of a public health impact than any other phase of care. Timely and efficient stroke treatment should be a priority for emergency department (ED) and prehospital providers. This article discusses the currently available and emerging treatment options in AIS focusing on the preservation of salvageable brain tissue, minimizing complications, and secondary prevention.

Patient evaluation overview

The initial evaluation of AIS should be focused on the efficient detection of functionally disabling neurologic deficits to optimize eligibility for time-dependent treatment options. A detailed discussion of AIS diagnosis is discussed elsewhere in this issue (See Lauren M. Nentwich’s article, “Diagnosis of Acute Ischemic Stroke,” in this issue). In short, an expedited neurologic examination should be performed including, but not limited to the National Institutes of Health Stroke Scale (NIHSS). Documentation of a NIHSS before stroke treatment and at the time of initial evaluation is a quality metric per The Joint Commission for Primary and Comprehensive Stroke Centers, which becomes the responsibility of the emergency medicine provider, unless neurologic expertise is available in house or via remote video telestroke services. Although formal NIHSS training and certification is not currently required per The Joint Commission, it is encouraged and freely available ( https://secure.trainingcampus.net/uas/modules/trees/windex.aspx?rx=nihss-english.trainingcampus.net ). Perhaps more important than a full NIHSS, at least initially, is to perform a brief stroke detection and severity screen, which can be performed in the ambulance or while being triaged in the ED. Prehospital stroke detection screens such as FAST (Facial drooping, Arm weakness, Speech difficulties and Time), CPSS (Cincinnati Prehospital Stroke Scale), LAPSS (Los Angeles Prehospital Stroke Screen), MASS (Massachusetts Stroke Scale), Med-PACS (Medic Prehospital Assessment for Code Stroke), OPSS (Ontario Prehospital Stroke Screening Tool), and ROSIER (Recognition of Stroke in the Emergency Room) have been linked with improved thrombolytic treatment rates and door-to-needle times. Severity scales such as the LAMS (Los Angeles Motor Scale), KPSS (Kurashiki Prehospital Stroke Scale), sNIHSS (Short NIHSS), CPSSS (Cincinnati Prehospital Stroke Severity Scale), VAN (vision, aphasia, neglect), and RACE (Rapid Arterial oCclusion Evaluation) have proven reasonably sensitive and specific tools to detect patients with emergent large vessel occlusion (ELVO) and may be used to trigger neurointerventional team activation, prehospital diversion, or interfacility transfer to a comprehensive stroke center.

Emphasis should be given to establishing the time the patient was “last known well,” that is, without symptoms, which is distinct from the time symptoms were first noted. The time last known well should be used in all cases as the equivalent of symptom onset unless the patient or witness is clearly able to recall the time symptoms began. This is important to ensure that symptom duration is not underestimated, resulting in treatment of the patient with thrombolytics beyond the approved treatment window.

It is also important to gain a sense of the patient’s premorbid functional status immediately before the stroke onset. This becomes important when weighing and discussing the risks and benefits of treatment options for reperfusion. This can usually be done by simply asking if the patient was functionally independent before stroke onset. A number of easy-to-use disability scales have been developed for this purpose, but the clinician often assesses a ‘gestalt’ version at the bedside. The modified Rankin Scale (mRS) is the most commonly used measure in stroke trials and measures global disability on a 6-point ordinal scale as shown in Table 1 .

Emergent and supportive care

Once the diagnosis of AIS has been made, it should go without saying that the basics of patient resuscitation, including the ABCs (airway, breathing, and circulation) take precedence, just as with any other ED patient. Especially important are temperature and glucose regulation, prevention of hypoxia, and optimization of blood pressure, and should remain priorities in all stroke patients. Cerebral perfusion pressure is the difference between the mean arterial pressure and intracranial pressure and is influenced by patient positioning and the progression of vasogenic edema in large territory infarcts. Unless the patient will, is, or has just received thrombolytic therapy, permissive hypertension (≤220 mm Hg systolic) should be allowed to promote cerebral autoregulation and perfusion of collateral vessels. Patients receiving thrombolytics should have their blood pressure maintained at or below 185/110 mm Hg in accordance with the American Heart Association/American Stroke Association (AHA/ASA) recommendations. Patients should be positioned with their heads in the midline position to promote optimal venous return and supine as tolerated, except in cases where impaired swallowing is noted, or there is concern for increased intracranial pressure or impaired cardiopulmonary function wherein the supine position may induce hypoxia. In these cases, elevating the head of the bed 15° to 30° is recommended. Hyperthermia (>38°C) is associated with poor neurologic outcomes and should be quickly corrected. The QASC (Quality in Acute Stroke Care) trial showed that hospital-wide supportive protocols focusing on the management of fever, hyperglycemia, and swallowing dysfunction were associated with improved 90-day functional outcomes (absolute difference, 15.7%; 95% confidence interval [CI], 5.8–25.4; P = .002) with a number needed to treat of 6. Empiric supplementary oxygen need not be administered unless it is required to maintain oxygen saturation of greater than 94%. Normoglycemia should be maintained, but the degree to which tight adherence is required remains unresolved.

Reperfusion with intravenous thrombolysis

The concept of thrombolysis in AIS is neither new nor without sustained controversy. Initial studies in the 1950s used streptokinase and urokinase, isolated from Streptococcus strains and human urine, respectively. Intracerebral hemorrhage (ICH) was a leading cause of death in these early investigations, which preceded computed tomography (CT) technology. In the 1960s, Meyer and colleagues used diagnostic angiography to perform investigations, first comparing intravenous (IV) plasmin with placebo, then combination therapy with streptokinase and heparin versus heparin alone. Although the former showed no benefit of the plasmin treated group, the latter showed greater mortality and ICH in the streptokinase-treated group. The MAST-E (Multicenter Acute Stroke Trial-Europe) and MAST-I (Multicenter Acute Stroke Trial-Italy) trials of streptokinase in the 1990s further confirmed increased risk of ICH and mortality, leading to the eventual abandonment of it as a treatment for AIS.

Concurrently, tissue plasminogen activator (t-PA) emerged as a clinically superior fibrinolytic to streptokinase in myocardial ischemia, which quickly translated into new investigations in AIS. This culminated in the landmark NINDS-II (National Institute of Neurological Disorders and Stroke) trial of 624 stroke patients published in 1995, which showed improved clinical outcome at 3 months for AIS patients treated with t-PA within 3 hours, despite a risk of symptomatic ICH of 6.4%, compared with 0.6% of controls ( P <.001). Earlier the same year, the ECASS (European Cooperative Acute Stroke Study)-I trial failed to demonstrate the same benefit among 620 patients randomized to t-PA versus placebo treated within 6 hours of symptom onset, which the authors attributed to a large number of protocol violations, accounting for 17.4% of the study population. Regardless, 1 year later the FDA approved t-PA for the treatment of AIS up to 3 hours after symptom onset.

The NINDS-II study has been criticized for having significant imbalances between treatment groups. However, the benefits of t-PA when given between 0 and 3 hours have been reinforced in post hoc analyses. Moreover, an independent commission concluded that there was no significant difference in treatment effect based on these imbalances. Overall, the NINDS trial found that t-PA–treated patients were 30% more likely to have minimal or no disability at 3 months for a number needed to treat of 7 to 8 for a favorable outcome.

Multiple subsequent studies in the following years either failed to demonstrate the treatment effect of the NINDS trial or were terminated owing to harm, including ECASS II and ATLANTIS (Alteplase Thrombolysis for Acute Noninterventional Therapy in Ischemic Stroke) Part A (0–6 hours), and ATLANTIS Part B (3–5 hours). The 0- to 3-hour treatment window remained the standard of care until publication of ECASS III in 2008, which showed clinical efficacy in treatment up to 4.5 hours (OR, 1.34; 95% CI, 1.02–1.76; P = .04). There was no difference in mortality, despite an increased risk of symptomatic ICH (7.9% for t-PA vs 3.5% for placebo; P <.001). The number needed to treat for a favorable outcome was 14 to 15. Unfortunately, these results were not replicated in the EPITHET (Echoplanar Imaging Thrombolytic Evaluation Trial) trial, which was published the same year and randomly assigned patients to t-PA or placebo with symptoms between 3 to 6 hours’ duration and a perfusion–diffusion mismatch on emergency MRI. Although EPITHET failed to achieve its primary endpoint (radiographic infarct growth at day 90), the t-PA–treated group was significantly associated with reperfusion, which correlated with improved neurologic outcomes.

The benefits of t-PA were further reinforced in 2009 in a metaanalysis inclusive of patients from ECASS 1, 2, and 3, and ATLANTIS showing an increased odds of a favorable outcome without difference in mortality (OR, 1.31; 95% CI, 1.10–1.56; P = .002). This led to the AHA/ASA revising its stroke guidelines accordingly the same year, recommending t-PA treatment up to 4.5 hours of symptom onset (class 1, level of evidence B). An updated pooled analysis was published in 2010 adding NINDS and EPITHET subjects totaling 3670 patients further supporting the time-dependent treatment effect with favorable outcomes up to 4.5 hours. In 2013, a joint recommendation by the AHA/ASA, American Academy of Neurology, American College of Emergency Physicians (ACEP), American Nurses Association, and Neurocritical Care Society supported the use of t-PA from 0 to 3 hours with a Level A recommendation and between 3 and 4.5 hours with a level B recommendation. The use of t-PA beyond 3 hours remains off-label in the United States, however, because it is FDA-approved only up to 3 hours of symptoms.

The IST-3 (Third International Stroke Trial) is the most recent large clinical trial to investigate the efficacy of IV t-PA up to 6 hours of symptom duration. The study had intended to enroll 6000 subjects in a 1:1 open, controlled design. However, owing to slow subject recruitment, the target was adjusted and in total, 3035 patients were enrolled. Because thrombolysis was already considered standard of care from 0 to 4.5 hours during the enrollment period, only patients for whom there remained clinical equipoise as to the efficacy of t-PA were recruited. Even among this population who were enrolled despite classical exclusion criteria and up to 6 hours after symptom onset, a significant ordinal reduction in disability was noted at 6 months. Most important, the subgroup of patients treated within 3 hours of onset showed a significant benefit of t-PA (OR, 1.64; 95% CI, 1.03–2.62). In a separate publication of 18-month follow-up of IST-3 subjects, the favorable shift in functional status persisted (OR, 1.3; 95% CI, 1.10–1.55; P = .002) and t-PA–treated patients self-reported better overall health ( P = .019). This trial is sometimes dismissed because it did not achieve its primary aim, namely, demonstration of improvement in the primary outcome, at the P <.05 level. It should be emphasized, however, that this is not the same as showing harm, simply that our confidence in the benefit of t-PA (outside of traditional NINDS criteria) is not as strong. When combined with the pool of all available trials of thrombolytics for stroke, IST-3 provided further evidence that t-PA treatment in stroke is not life saving, but rather autonomy preserving.

Although subject to methodologic heterogeneity, pooled analyses do provide valuable insight as to the overall efficacy and safety of IV t-PA. Systemic thrombolysis has, for some time, been considered the worldwide standard of care in AIS; thus, the replication of the landmark clinical trials that have been subject to heavy criticism would be unethical given a universal lack of clinical equipoise. The most robust of the pooled analyses performed to date is a Cochrane review published in 2014 inclusive of 10,187 patients from 27 different clinical trials. Of the subjects in this pooled analysis, 7012 were randomized to IV t-PA versus placebo, whereas the others involved other thrombolytics. The key findings reinforced the individual studies’ conclusions, namely that “time is brain” and faster treatment is better. Not surprisingly, dichotomized treatment from 0 to 3 hours outperformed treatment between 3 and 6 hours (OR, 1.56 [95% CI, 1.28–1.90] vs OR 1.07 [95% CI 0.96–1.2]). The authors’ conclusion that treatment with t-PA improves outcomes for patients if given up to 4.5 hours after stroke onset has been criticized and led to some debate. Although heterogeneous in total, a homogeneous subgroup of 1779 patients across 6 trials showed clear benefit of t-PA when given early, within the first 3 hours (OR, 0.65; 95% CI, 0.54–0.80; P <.0001). Moreover, an individual patient data metaanalysis from 6756 patients conducted in 2014 by the Stroke Thrombolysis Trialists’ Collaboration in 2014 concluded that the benefit of t-PA extends to sometime beyond 4.5 hours. Ninety-day mortality was 1.4% higher among those receiving t-PA (hazard ratio, 1.11; 95% CI, 0.99–1.25; P = .07), but was offset by an absolute increase in disability-free survival of 10% if treated within 0 to 3 hours or 5% if treated within 3 to 4.5 hours.

Nevertheless, the methodology and conclusions of recent pooled analyses have been called into question and has resulted in the recent revised grading of evidence supporting the use of t-PA by both the ACEP and the Canadian Association of Emergency Physicians. Whereas ACEP supports the use of t-PA up to 4.5 hours (level B recommendation), the Canadian Association of Emergency Physicians has issued a conditional recommendation against the use of t-PA in the 3- to 4.5-hour window, largely owing to the increased risk of ICH and possible increase in 90-day mortality until further research is available.

Clearly, the lack of consensus regarding the use of t-PA continues to be the subject of debates, extending to the highest levels of health care policy. What should not be lost in these debates, however, is the patient’s preference. Although the risk of ICH is relatively small, it is by no means insignificant (4.9%–7.7%) and the decision to risk a potentially expedited death for a reduced likelihood of long-term disability is a highly personal one. Hence, a brief discussion of risks and benefits, and, when feasible, the use of shared-decision making between the patient/surrogate and health care provider is recommended (ACEP level C recommendation).

The issue of informed consent and informed refusal

Whether informed consent for IV t-PA is needed in acute stroke is medicolegally and ethically unclear, and has been the source of a fair amount of debate. Because it has been established that “time is brain,” deferring treatment until risks and benefits have been discussed adequately could be viewed by some as withholding the established standard of care. Additionally, the acute stress of the event on the patient and family often impedes comprehension and may quickly lead to paternalistic decision making. Although t-PA use in ischemic stroke is not accepted universally as an emergency exception to informed consent, a survey in JAMA suggests that overall rates favoring t-PA treatment in stroke is similar to rates for cardiopulmonary resuscitation for cardiac arrest. Yet, the unique risk/benefit profile of t-PA draws no parallel or precedent in medicine and so general consensus among the stroke community favors obtaining informed consent. The use of a visual aid such as the one proposed by Gadhia and colleagues may facilitate a balanced and efficient discussion. Implied consent may be invoked per an institutional standard of care when the patient lacks decision-making capacity and no surrogate can be located after a reasonable effort.

Because t-PA is considered standard of care in certain circumstances, it has been proposed that the risks/benefits discussion be focused on informed refusal rather than informed consent. This model preserves patient autonomy and the right of self-determination, while emphasizing the substantial, yet time-dependent benefit of t-PA. Patient-centered decision support tools aimed to streamline a brief summary of risks/benefits have been developed, and, among focus groups of stroke survivors, caregivers, emergency physicians, and advanced practice nurses are well-received when framed positively and displayed as simplified graphs. The Rapid Evaluation for Stroke Outcomes using Lytics in a Vascular Event (RESOLVE) decision aid is currently being tested and consists of a 3-page patient tool and single page clinician tool summarizing individualized risk prediction models.

Safety of tissue plasminogen activator in stroke mimics

The emphasis on shortening door-to-needle times to expedite stroke thrombolysis comes at some expense to the specificity of stroke diagnosis. Certainly, some of the stroke mimics like hemiplegic migraine and Todd’s paralysis may be indistinguishable from stroke in the early moments of care. In a recent metaanalysis of reported case series inclusive of 8942 total patients treated with t-PA, 392 were found to be stroke mimics. Symptomatic ICH rates (defined as imaging evidence of hemorrhage as well as a NIHSS increase of ≥4 points) were 0.5% (95% CI, 0%-2%) and the risk was significantly lower in stroke mimics (risk ratio 0.33; 95% CI, 0.14–0.77; P = .010). The rate of orolingual edema was also low (0.3%; 95% CI, 0%-2%) and favorable outcomes were 3-fold higher among stroke mimics. From a medicolegal perspective, a systematic review of malpractice litigation cases in AIS revealed that of 40 cases related to IV t-PA, only 2 (5%) were related to complications of administration, whereas 38 (95%) were related to failure to administer.

Alternative and emerging thrombolytic agents

Alteplase remains the thrombolytic of choice in clinical practice, but is not the ideal agent. Its short half-life (5 minutes) requires a continuous infusion after initial bolus and its fibrin selectivity is moderate on the spectrum of thrombolytic agents. Its limited efficacy in causing reperfusion, particularly in large vessels, as well as its associated risk of ICH and postulated neurotoxicity has kept researchers searching for newer and better thrombolytics. Tenecteplase (TNK) is a genetically engineered alternative to alteplase that has more than twice the half-life and is 15 times more fibrin specific. It also is more resistant to plasminogen activator inhibition, which is an endogenous counteraction to thrombolytic effect, and is associated with less degradation of apolipoprotein A-1, the main protein component of high-density lipoprotein. For these reasons, it is currently the preferred lytic in patients with ST-elevation myocardial infarction. TNK may be even more appealing in AIS because it causes less hypofibrinogenemia, and early degradation of fibrinogen is associated with ICH. Three multicenter RCTs comparing TNK to alteplase are currently underway (TEMPO-2 [A Randomized Controlled Trial of TNK-tPA Versus Standard of Care for Minor Ischemic Stroke With Proven Occlusion], NOR-TEST [Norwegian Tenecteplase Stroke Trial], and EXTEND-IA TNK [Tenecteplase Versus Alteplase Before Endovascular Therapy for Ischemic Stroke]), but multiple phase II trials have shown promise and no increase in risk for symptomatic ICH. One study found that TNK at a dose of 0.25 mg/kg was superior to alteplase in reperfusion and clinical outcome at 90 days.

Desmoteplase, a recombinant fibrinolytic isolated from the saliva of the vampire bat, has an even longer half-life and is more fibrin selective than either TNK or alteplase. In fact, desmoteplase activity was found to be 105,000 times higher in the presence of fibrin than without, compared with alteplase, which showed only a 550-fold increase. Phase II clinical trials (DEDAS [Dose Escalation of Desmoteplase for Acute Ischemic Stroke ]/DIAS [Desmoteplase in Acute Ischemic Stroke Trial]) suggested safety and efficacy of desmoteplase in stroke patients treated 3 to 9 hours from symptom onset, but these findings were not replicated in the subsequent DIAS 2 or 3 trials. However, despite a median time to treatment of 7 hours in DIAS 3, there were no safety concerns and symptomatic ICH rates were low (3%) and similar to placebo. Although these trials were unable to extend the therapeutic window of systemic thrombolysis, desmoteplase remains an appealing alternative thrombolytic agent. Several other agents, including retaplase and microplasmin, are showing promise in the preclinical phase.

Interventional stroke treatment

The history of endovascular stroke treatment began with intraarterial (IA) administration of fibrinolytic agents. In 1999, the PROACT (Intra-arterial Prourokinase for Acute Ischemic Stroke) II trial showed improved functional outcomes at 90 days in patients treated with IA prourokinase within 6 hours of onset of large middle cerebral artery occlusion ( P = .04). Catheter-directed thrombolysis offers the theoretic advantage of targeted drug delivery directly at the site of the thrombus to achieve recanalization at a lesser total drug dose. Although prohibited in the PROACT II trial, in clinical practice recanalization was further augmented through guidewire-mediated mechanical clot disruption during IA thrombolysis.

Mechanical clot retrieval devices were first developed to recover displaced coils from endovascular procedures. The first device to gain FDA approval was the MERCI retriever in August 2004, after the device was shown to successfully recanalize 43% of large vessel occlusions when used alone, or 64% in combination with IA t-PA. Suction thrombectomy using an aspiration device such as the Penumbra System gained FDA approval in 2007 after showing recanalization in 82% of patients in a prospective, single-arm study of 125 subjects. It should be noted, however, that FDA approval was granted solely based on the ability of these devices to successfully recanalize thrombosed vessels in acute stroke patients, not from efficacy in improving functional outcomes.

The initial randomized, controlled trials investigating the efficacy of interventional stroke treatment on functional outcomes included 3 studies: IMS (Interventional Management of Stroke) III, MR RESCUE (Mechanical Retrieval and Recanalization of Stroke Clots Using Embolectomy) and SYNTHESIS (Local Versus Systemic Thrombolysis for Acute Ischemic Stroke), all published in the same issue of the New England Journal of Medicine on February 8, 2013. None of the studies showed a statistically significant difference in clinical outcome of patients treated with endovascular therapy (EVT) versus systemic thrombolysis alone. SYNTHESIS remains the only trial to date that directly compared EVT without IV t-PA with t-PA alone. 362 AIS patients within 4.5 hours of onset were randomized to IV t-PA or EVT (IA or clot disruption/retrieval, or both) and the proportion alive without disability (mRS 0–1) at 3 months were similar between groups (30.4% vs 34.8%; adjusted OR 0.71; 95% CI, 0.44–1.14; P = .16). The rates of symptomatic ICH were similar between groups (∼6%).

IMS III and MR RESCUE were similar in comparing IV t-PA alone with IV t-PA + EVT. MR RESCUE used perfusion-weighted imaging (CT or MRI) to differentiate patients thought to have a “favorable” imaging pattern, in whom a greater portion of potentially salvageable tissue was present. A “favorable” imaging pattern was defined as the infarct core was 90 mL or less and the proportion of predicted infarct was 70% or less of the at-risk region. Patients were randomized to EVT + IV t-PA or IV t-PA alone, regardless of the imaging pattern and embolectomy was not found to be superior in reducing 90-day mRS scores among penumbral ( P = .23) or nonpenumbral ( P = .32) imagining patterned patients. IMS III, on the other hand, did not require perfusion imaging or even vessel imaging showing evidence of large vessel occlusion before randomization. In this study, stroke patients who had received IV t-PA within 3 hours of onset were randomized 2:1 to EVT (mechanical thrombectomy or IA) or no adjunctive therapy. The study was terminated prematurely owing to futility after 656 of an anticipated 900 patients were enrolled. The primary outcome measure was the proportion of patients with a mRS of 0 to 2 at 90 days, which did not differ between groups (40.8% for EVT vs 38.7% for t-PA alone; absolute adjusted difference, 1.5%; 95% CI, -6.1–9.1). Mortality and symptomatic ICH rates were similar as well. A trend toward a favorable outcome was noted in patients with a carotid T-type or L-type occlusion of the internal carotid artery, M1/A1 or tandem internal carotid artery and M1 occlusion in a subgroup analysis of 306 subjects (47%) with baseline CT or MR angiographic data. This finding may have contributed to baseline vessel imaging becoming standard in all subsequent EVT trials.

Collectively, the negative results of IMS III, SYNTHESIS, and MR RESCUE helped to reestablish clinical equipoise to a stroke treatment strategy that had become pervasive without any compelling outcome data from a randomized, controlled trial. This set the stage for a new round of EVT trials, which benefitted from the advent of third-generation devices, called stent retrievers or “stent-trievers,” which were felt to more consistently, efficiently, and safely establish recanalization. These devices include the SOLITAIRE and TREVO devices, which received FDA approval in 2012. Simultaneously, stroke systems of care continued to evolve, thereby allowing for more rapid identification of ELVO strokes. These advances helped to fuel the second round of clinical trials that would prove revolutionary to acute stroke treatment.

The second wave of EVT trials came in early 2015 with 5 prospective, randomized multicenter studies from around the world demonstrating benefit of EVT over IV t-PA alone. These include MR CLEAN (Multicenter Randomized Clinical trial of Endovascular treatment for Acute ischemic stroke in the Netherlands), EXTEND-IA, ESCAPE (Endovascular Treatment for Small Core and Proximal Occlusion Ischemic Stroke), SWIFT PRIME (Solitaire With the Intention For Thrombectomy as PRIMary Endovascular Treatment), and REVASCAT (Endovascular Revascularization With Solitaire Device Versus Best Medical Therapy in Anterior Circulation Stroke Within 8 Hours). Partial results of 2 additional studies, THRACE (Trial and Cost Effectiveness Evaluation of Intra-arterial Thrombectomy in Acute Ischemic Stroke) and THERAPY (Assess the PENUMBRA System in the Treatment of Acute Stroke) have also been reported, but remain in prepublication phase. Despite variations in inclusion/exclusion criteria across studies, all 5 shared several similarities, most notably the confirmation of ELVO via vascular imaging acquisition (CTA or MRA) at baseline, as well as a common outcome measure, the mRS at 90 days. A favorable outcome was considered an mRS of 0 to 2. Of note, only the first trial to report its results, MR CLEAN, completed enrollment; all other trials were halted prematurely owing to meeting predetermined efficacy endpoint or owing to lack of equipoise in light of other positive trials.

The MR CLEAN study concluded in March 2014 after enrolling 500 subjects across 16 sites in the Netherlands. The Dutch Ministry of Health was a motivating factor in trial completion by provisionally reimbursing for EVT only in the setting of this trial on the heels of the prior negative studies. Patients were enrolled within 6 hours of symptom onset if NIHSS was 2 or greater. Of the 233 patients in the interventional arm, 33% had a favorable outcome compared with 19% of the 267 in the control group. This was an absolute difference of 14% (adjusted OR, 1.67; 95% CI, 1.2–2.3).

The ESCAPE study was conducted primarily in Canada, including 22 sites that enrolled 316 of an intended 500 subjects. Patients with an NIHSS of 6 or greater were included up to 12 hours after symptom onset. Of the 165 patients in the intervention arm, 53% had a favorable outcome compared with 29% in the control arm, for an absolute difference of 24% ( P <.001).

The EXTEND-IA trial was conducted in Australia and New Zealand across 14 different sites. Only 70 of the planned 100 patients (35 intervention, 35 control) were enrolled. Patients with an NIHSS of 2 or greater were included up to 6 hours from symptom onset. Although the 90-day mRS was an outcome measure, this was a secondary endpoint in this study, and the coprimary endpoints were radiographic reperfusion at 24 hours and early neurologic recovery (improvement of NIHSS by ≥8 points). The absolute difference in favorable outcomes at 90 days between treatment groups was 31% favoring EVT (71% vs 40%; P = .009).

The SWIFT PRIME trial enrolled only 196 of an intended 833 patients across 39 sites, primarily in the United States, before being halted owing to efficacy. This trial used a cutoff of 8 on the NIHSS and enrolled up to 6 hours from symptom onset. All patients received baseline CT/CTA/CT perfusion or diffusion-weighted MRI for imaging-based selection. Of the 98 in the intervention arm, 60% achieved a favorable outcome, compared with 36% of the controls, for an absolute difference of 24% ( P <.001).

The REVASCAT trial was the last to reports results, halting enrollment after only 206 of an intended 690 patients owing to loss of equipoise after the release of the preceding trials’ data. The NIHSS cutoff was 6 and patients were included up to 8 hours from symptom onset. The absolute difference between treatment groups in terms of a good outcome at 90 days was 16% favoring EVT (44% vs 28%; P <.001).

Clearly there now exists robust evidence supporting the use of EVT in patients with ELVO on baseline vessel imaging. The timeframe for initiation of EVT ranged from 6 to 12 hours in these studies, but evidence suggests that the effectiveness decreases over time. Of note, the trials with the fastest time to recanalization showed the largest treatment effect (EXTEND-IA, SWIFT PRIME, and ESCAPE). Furthermore, the “picture-to-puncture” times (ie, the time from first vascular image to groin puncture) was less than 60 minutes in ESCAPE and SWIFT PRIME, which would be nearly impossible to replicate without optimally efficient ED and neurointerventionalist activation processes. In the short time since publication of the 5 positive EVT trials, supporting guidelines from the AHA/ASA and Society of Neurointerventional Surgeons have already followed suit. EVT should be considered the standard of care for patients with ELVO, within 6 hours of symptom onset, and regional systems-of-care initiatives should focus on ensuring timely access to this treatment for as many patients as possible.

Processes to improve the availability and speed of access to endovascular therapies should be regional and institution specific. Some examples include the administration of t-PA in the CT scanner bay during dedicated vessel image acquisition, as well as EMS destination protocols based on prehospital stroke severity scale use to improve the proportion of ELVOs triaged to endovascular capable stroke centers. Mobile stroke units, ambulances equipped with a portable CT scanner, point-of-care laboratory capabilities, and access to stroke expertise are becoming increasingly common, have CT angiography capabilities, and have been shown to reduce time to EVT.