Guidelines for the Early Management of Patients With Acute Ischemic Stroke
A Guideline for Healthcare Professionals From the American Heart Association/American Stroke Association
Background and Purpose—The authors present an overview of the current evidence and management recommendations for evaluation and treatment of adults with acute ischemic stroke. The intended audiences are prehospital care providers, physicians, allied health professionals, and hospital administrators responsible for the care of acute ischemic stroke patients within the first 48 hours from stroke onset. These guidelines supersede the prior 2007 guidelines and 2009 updates.
Methods—Members of the writing committee were appointed by the American Stroke Association Stroke Council’s Scientific Statement Oversight Committee, representing various areas of medical expertise. Strict adherence to the American Heart Association conflict of interest policy was maintained throughout the consensus process. Panel members were assigned topics relevant to their areas of expertise, reviewed the stroke literature with emphasis on publications since the prior guidelines, and drafted recommendations in accordance with the American Heart Association Stroke Council’s Level of Evidence grading algorithm.
Results—The goal of these guidelines is to limit the morbidity and mortality associated with stroke. The guidelines support the overarching concept of stroke systems of care and detail aspects of stroke care from patient recognition; emergency medical services activation, transport, and triage; through the initial hours in the emergency department and stroke unit. The guideline discusses early stroke evaluation and general medical care, as well as ischemic stroke, specific interventions such as reperfusion strategies, and general physiological optimization for cerebral resuscitation.
Conclusions—Because many of the recommendations are based on limited data, additional research on treatment of acute ischemic stroke remains urgently needed.
- AHA Scientific Statements
- acute cerebral infarction
- emergency medical services
- tissue plasminogen activator
Despite the increase in the global burden of stroke, advances are being made. In 2008, after years of being the third-leading cause of death in the United States, stroke dropped to fourth.1 In part, this may reflect the results of a commitment made by the American Heart Association/American Stroke Association (AHA/ASA) more than a decade ago to reduce stroke, coronary heart disease, and cardiovascular risk by 25% by the year 2010 (a goal met a year early in 2009). The reason for the success was multifactorial and included improved prevention and improved care within the first hours of acute stroke. To continue these encouraging trends, the public and healthcare professionals must remain vigilant and committed to improving overall stroke care. This document addresses opportunities for optimal stroke care in the acute phase of the ischemic stroke.
The intended audience of these updated guidelines is healthcare professionals involved in the emergency identification, evaluation, transport, and management of patients with acute ischemic stroke. This includes prehospital care providers, emergency department (ED) physicians and nurses, stroke team members, inpatient nurses, hospitalists, general medicine physicians, hospital administrators, and ancillary healthcare personnel. These guidelines deal with the acute diagnosis, stabilization, and acute medical and surgical treatments of acute ischemic stroke, as well as early inpatient management, secondary prevention, and complication management. Over the past several years, several new guidelines, policy statements, and recommendations on implementation strategies for emergency medical services (EMS) within stroke systems of care, imaging in acute ischemic stroke, management of stroke in infants and children, nursing and interdisciplinary care in acute stroke, primary prevention of ischemic stroke, stroke systems of care, and management of transient ischemic attack (TIA) related to acute ischemic stroke have been published by the AHA/ASA. To minimize redundancy, the reader will be referred to these publications where appropriate.2–10
The Stroke Council of the AHA/ASA commissioned the assembled authors, representing the fields of cardiology, emergency medicine, neurosurgery, nursing, radiology, rehabilitation, neurocritical care, endovascular neurosurgical radiology, and vascular neurology, to completely revise and update the guidelines for the management of acute ischemic stroke.11–13 In writing these guidelines, the panel applied the rules of evidence and the formulation of strength of recommendations used by other panels of the AHA/ASA (Tables 1 and 2). The data were collected through a systematic review of the literature. Because of the wide scope of the guidelines, individual members of the panel were assigned as primary and secondary authors for individual sections, then the panel assessed the complete guidelines. If the panel concluded that data supported or did not support the use of a specific intervention, appropriate recommendations were made. In some instances, supporting evidence based on clinical trial research was not available for a specific intervention, but the panel has made a specific recommendation on the basis of pathophysiological reasoning and expert practice experience. In cases in which strong trial, physiological, and practice experience data were not available, no specific recommendation was made. Recommendations that have been changed or added since the publication of the previous guideline are accompanied by explicit statements indicating the revised or new status.
This publication serves as a current comprehensive guideline statement on the management of patients with acute ischemic stroke. This publication supersedes prior guidelines and practice advisories published by the AHA/ASA relevant to acute ischemic stroke.11–14 The reader is also encouraged to read complementary AHA/ASA articles, including statements on the development of stroke systems of care, EMS integration in stroke systems, telemedicine, and neuroimaging in acute stroke, which contain more detailed discussions of several aspects of acute stroke management.2–5
This document uses a framework based on the AHA stroke systems of care publication by Schwamm et al4 to provide a framework of how to develop stroke care within a regional network of healthcare facilities that provide a range of stroke care capabilities. Similarly, for an individual patient, this document draws on the 2010 advanced cardiac life support stroke chain of survival15 (Table 3), which describes the critical links to the process of moving a patient from stroke ictus through recognition, transport, triage, early diagnosis and treatment, and the final hospital disposition. Within regions and institutions, the exact composition of the system and chain may vary, but the principles remain constant: preparation, integration, and an emphasis on timeliness.
Public Stroke Education
The chain of events favoring good functional outcome from an acute ischemic stroke begins with the recognition of stroke when it occurs. Data show that the public’s knowledge of stroke warning signs remains poor.16 Fewer than half of 9-1-1 calls for stroke events were made within 1 hour of symptom onset, and fewer than half of those callers thought stroke was the cause of their symptoms.17 Many studies have demonstrated that intense and ongoing public education about the signs and symptoms of stroke improves stroke recognition.18 The California Acute Stroke Pilot Registry (CASPR) reported that the expected overall rate of fibrinolytic treatment within 3 hours could be increased from 4.3% to 28.6% if all patients arrived early after onset, which indicates a need to conduct campaigns that educate patients to seek treatment sooner.19 Effective community education tools include printed material, audiovisual programs, lectures, and television and billboard advertisements.20 Stroke education should target not only prospective patients but also their family members and caregivers, empowering them to activate the emergency medical system. Stroke education campaigns have been successful among elementary and middle school students.21,22
Before 2008, the 5 “Suddens” of stroke warning signs (sudden weakness; sudden speech difficulty; sudden visual loss; sudden dizziness; sudden, severe headache) were used widely in public education campaigns. The FAST (face, arm, speech, time) message campaign, first promoted a decade ago, is being reintroduced in public education efforts. One or more of face weakness, arm weakness, and speech difficulty symptoms are present in 88% of all strokes and TIAs.23 In one study, 100% of lay individuals remembered 3 months after education that facial droop and slurred speech are stroke warning signs, and 98% recalled arm weakness or numbness.24 Regardless of the message, effective public education requires repetition for a sustained impact.
Another central public education point is the message to call 9-1-1 promptly when a stroke is suspected. Despite a decade of stressing the role of 9-1-1 and EMS in stroke, the recent National Hospital Ambulatory Medical Care Survey (NHAMCS) showed that only 53% of stroke patients used EMS.25 Multiple studies have reported the benefits of 9-1-1 use and EMS involvement in acute stroke. Prehospital delays are shorter and initial computed tomography (CT) or magnetic resonance imaging (MRI) scans are obtained sooner if stroke patients are transported by ambulance.25 Advance notification of stroke patient arrival by EMS also shortens the time to be seen for initial evaluation by an emergency physician, shortens the time to brain imaging, and increases the use of the intravenous recombinant tissue-type plasminogen activator (rtPA) alteplase.26
Prehospital Stroke Management
After the 2007 publication of the “Guidelines for the Early Management of Adults With Ischemic Stroke,”13 the AHA/ASA published a policy statement, “Implementation Strategies for Emergency Medical Services Within Stroke Systems of Care,” from the Expert Panel on Emergency Medical Services Systems and the Stroke Council.5 This statement serves as the blueprint that defines the critical roles of EMS and EMS systems (EMSS) in optimizing stroke care. EMS refers to the full scope of prehospital stroke care, including 9-1-1 activation and dispatch, emergency medical response, triage and stabilization in the field, and ground or air ambulance transport; EMSS refers to the system that involves the organization of public and private resources and includes the community, emergency healthcare personnel, public safety agencies, emergency facilities, and critical care units. Issues related to communication, transportation, access to care, patient transfer, mutual aid, and system review and evaluation are addressed in EMSS. To reach full potential, stroke systems of care must incorporate EMSS into the process.
The “Implementation Strategies for Emergency Medical Services Within Stroke Systems of Care” policy statement outlines specific parameters that measure the quality of an EMSS, including the following:
Stroke patients are dispatched at the highest level of care available in the shortest time possible.
The time between the receipt of the call and the dispatch of the response team is <90 seconds.
EMSS response time is <8 minutes (time elapsed from the receipt of the call by the dispatch entity to the arrival on the scene of a properly equipped and staffed ambulance).
Dispatch time is <1 minute.
Turnout time (from when a call is received to the unit being en route) is <1 minute.
The on-scene time is <15 minutes (barring extenuating circumstances such as extrication difficulties).
Travel time is equivalent to trauma or acute myocardial infarction calls.5
With the use of electronic EMS data capture and storage, these performance measures are readily available for review and system improvement.
The call to the 9-1-1 dispatcher is the first link in the stroke chain of survival.15 To facilitate the recognition of stroke and provide adequate prehospital stroke care by EMS, statewide standardization of telecommunication programs, stroke education modules, and care protocols is recommended.27–29 The provision of ongoing education to dispatchers will improve their skills in recognizing the signs and symptoms of stroke.30 In one study, 9-1-1 dispatchers correctly identified 80% of all stroke calls if the caller mentioned specific words such as stroke, facial droop, weakness/fall, or communication problems.31 If there is diagnostic concordance of stroke between dispatchers and paramedics, the scene time and run times are shortened.32 Once a stroke is suspected, it becomes a high-priority dispatch.
EMS Assessment and Management
As detailed in the recent update of the AHA’s Emergency Cardiovascular Care Committee recommendations for acute stroke, the primary goals of EMS assessment and management are rapid evaluation, early stabilization, neurological evaluation, and rapid transport and triage to a stroke-ready hospital.15 As in all scene responses, EMS personnel must assess and manage the patient’s airway, breathing, and circulation (ABCs). Most patients with acute ischemic stroke do not require emergency airway management or acute interventions for respiratory and circulatory support.
Several prehospital interventions to improve the overall physiological state may be beneficial to patients with suspected acute stroke. Prehospital care has emerged from general principles of resuscitation. Although data from prehospital clinical trials are not always stroke-specific, they do provide guidance for making recommendations for potential stroke patients. Although the routine use of supplemental oxygen remains unproven, supplemental oxygen to maintain oxygen saturations >94% is recommended after cardiac arrest and is reasonable for patients with suspected stroke.15,33 In potential stroke patients who are hypotensive, defined as blood pressure significantly lower than premorbid state or systolic blood pressure <120 mm Hg, placement of the head of the stretcher flat and administration of isotonic saline may improve their cerebral perfusion. In contrast, in patients who are hypertensive (systolic blood pressure ≥140 mm Hg), the benefit of routine prehospital blood pressure intervention is not proven; consultation with medical control may assist in making treatment decisions regarding patients with extreme hypertension (systolic blood pressure ≥220 mm Hg). The types of antihypertensive medications used in this setting are described in the inpatient section of hypertension management. Hypoglycemia is frequently found in patients with strokelike symptoms; thus, prehospital glucose testing is critical. If a patient is found to have blood glucose levels <60 mg/dL, intravenous administration of glucose may resolve the neurological deficits. For nonhypoglycemic patients, excessive dextrose-containing fluids have the potential to exacerbate cerebral injury; thus, normal saline is more appropriate if rehydration is required. Lastly, establishment of an intravenous line in the field not only facilitates the administration of prehospital medications and fluids but can also shorten treatment times in the ED. When possible, EMS may obtain blood samples for laboratory testing en route to the ED, where they can immediately be given to the laboratory on arrival. These steps may take place while stroke patients are being transported. There should be no delay in getting the stroke patient to the ED by establishing intravenous access, checking blood glucose level, or obtaining blood samples. Although all of these recommendations represent the ideal scenario, it is critical that interventions not delay transport of the patient to the hospital.
Once the initial patient assessment and stabilization are complete, EMS personnel may obtain a focused history from the patient or bystanders. The most important piece of information necessary for potential fibrinolytic treatment is the time of symptom onset, defined as the time the patient was last known normal. Often patients are aphasic or are unaware of their deficits and arrive without accompanying family who can provide necessary information. Thus, it is critical for EMS personnel to establish the time the patient was last known normal from those at the scene. Other important historical elements include any sign of seizure activity or trauma before onset of symptoms. Elements of the past medical history can assist in the prehospital diagnosis of stroke or a stroke mimic, such as history of seizures or hypoglycemia. A history of prior stroke, diabetes mellitus, hypertension, and atrial fibrillation all increase the likelihood that the patient’s symptoms are caused by stroke. EMS personnel can identify current medications, especially any anticoagulants, and recent illnesses, surgery, or trauma. EMS personnel also can obtain phone numbers at which family members or witnesses can be reached by ED personnel to provide further history after arrival. When stroke patients are unable to provide information to hospital care providers, EMS personnel may consider transporting a family member along with the patient.
Once the primary survey is complete, EMS personnel should perform a more focused organ system assessment, but transport should not be delayed. Numerous prehospital neurological assessment tools have been developed to accurately identify stroke patients, which facilitates appropriate field treatment, prearrival notification, and routing to an appropriate hospital destination.34,35 Given regional differences in stroke systems of care, local EMS personnel may use a regionally appropriate, validated prehospital neurological assessment tool. As with all prehospital evaluations, EMS personnel typically complete a secondary survey, reviewing the head and neck for signs of trauma, auscultating the heart and lungs, and observing the patient’s extremities for any signs of trauma. To ensure optimal prehospital care, hospital stroke providers should provide feedback to EMS agencies as part of continuous quality improvement projects.
As is the case for patients with trauma or acute myocardial infarction, prehospital notification by EMS of a potential stroke is essential. Several studies have shown that prehospital notification leads to significant reductions in several stroke time benchmarks, including time from arrival to physician assessment, CT performance, and CT interpretation, and is associated with higher rates of intravenous rtPA administration.26,36–38
Air Medical Transport
Air transport service is particularly useful to facilitate stroke care in remote areas. As part of regional stroke systems of care, activation of air medical transport for stroke is reasonable when ground transport to the nearest stroke-capable hospital is >1 hour.5 Local stroke hospitals may provide expertise to help create activation protocols and in-flight stroke management protocols to ensure safe and appropriate patient transports.39,40
With the development of primary stroke centers (PSCs) and comprehensive stroke centers (CSCs), which offer intra-arterial strategies, interhospital transfers of acute stroke patients are increasingly common. Some patients are transferred before fibrinolytic therapy, whereas others receive intravenous rtPA and then are transferred for higher-level care. Delaying intravenous rtPA therapy until after transport in otherwise eligible patients decreases the chance for a good outcome. In the “drip-and-ship” model, in which the patient begins to receive standard-dose intravenous rtPA before transfer, well-designed protocols that include strict adherence to blood pressure guidelines, assessment for clinical deterioration and bleeding, and aspiration precautions ensure safe interhospital transport. Transport personnel should be able to contact medical command or the receiving facility about any change in the patient’s condition en route.
Conclusions and Recommendations
EMSS are essential elements in all stroke systems of care. Beginning with public education on recognizing signs and symptoms of stroke and the need for calling 9-1-1, these first elements in the stroke chain of survival are arguably the most important. Calling 9-1-1 and using EMS are the preferred ways of providing optimal prehospital stroke care and transport to stroke centers. Specific time frames have been established for the EMSS to follow on dispatch, response, and on-scene activities, and this should be monitored continuously. Notification of the receiving institution before arrival is critical because it facilitates the rapid diagnosis and management of stroke patients. All efforts must be made to avoid unnecessary delays during patient transport. Statewide, standardized EMS education and stroke care protocols for EMSS improve prehospital stroke recognition and management.
To increase both the number of patients who are treated and the quality of care, educational stroke programs for physicians, hospital personnel, and EMS personnel are recommended (Class I; Level of Evidence B). (Unchanged from the previous guideline13)
Activation of the 9-1-1 system by patients or other members of the public is strongly recommended (Class I; Level of Evidence B). 9-1-1 Dispatchers should make stroke a priority dispatch, and transport times should be minimized. (Unchanged from the previous guideline13)
Prehospital care providers should use prehospital stroke assessment tools, such as the Los Angeles Prehospital Stroke Screen or Cincinnati Prehospital Stroke Scale (Class I; Level of Evidence B). (Unchanged from the previous guideline13)
EMS personnel should begin the initial management of stroke in the field, as outlined in Table 4 (Class I; Level of Evidence B). Development of a stroke protocol to be used by EMS personnel is strongly encouraged. (Unchanged from the previous guideline13)
Patients should be transported rapidly to the closest available certified PSC or CSC or, if no such centers exist, the most appropriate institution that provides emergency stroke care as described in the statement (Class I; Level of Evidence A). In some instances, this may involve air medical transport and hospital bypass. (Revised from the previous guideline13)
EMS personnel should provide prehospital notification to the receiving hospital that a potential stroke patient is en route so that the appropriate hospital resources may be mobilized before patient arrival (Class I; Level of Evidence B). (Revised from the previous guideline13)
Designation of Stroke Centers and Stroke Care Quality Improvement Process
Stroke Systems of Care
The ASA task force on the development of stroke systems has defined key components of a regional stroke system of care and recommended methods for the implementation of stroke systems.4 Stroke systems of care integrate regional stroke facilities, including acute stroke-ready hospitals (ASRHs) that often have telemedicine and teleradiology capability, primary and comprehensive stroke centers, EMSS, and public and governmental agencies and resources. The goals of creating stroke systems of care include stroke prevention, community stroke education, optimal use of EMS, effective acute and subacute stroke care, rehabilitation, and performance review of stroke care delivery. Essential to effective stroke systems of care are hospitals with the capacity and commitment to deliver acute stroke care, both in the ED and on the stroke unit. In regions with effective stroke systems, the majority of patients are now being transported to these stroke centers, which optimizes their chances for timely appropriate therapy and admission to stroke units, both of which decrease the morbidity and mortality associated with stroke.41,42
Hospital Stroke Capabilities
Primary Stroke Center
The definition of a PSC was first published in 2000.43 This article defined the critical prehospital and hospital elements to deliver effective and efficient stroke care. Since The Joint Commission (TJC) began providing PSC certification in 2004, >800 certified PSCs have been established in the United States (as of January 2011).44 Regardless of certifying agent (TJC or state health department), it is mandatory for all PSCs to closely track their performance on key quality stroke care measurements. In cluster controlled clinical trials comparing patient outcomes in PSCs with those in community hospitals without specialized stroke care, patients with ischemic stroke treated in centers with dedicated stroke resources had better clinical outcomes45 and increased rates of intravenous rtPA administration.20 In addition, numerous observational studies have demonstrated that PSC certification improves stroke care in many ways, for instance, by shortening door to physician contact time, door to CT time, and door to intravenous rtPA time, as well as by increasing rates of intravenous rtPA use.46–48 Hospitals that have implemented organized stroke care have demonstrated sustained improvements in multiple measures of stroke care quality, including increased use of intravenous rtPA, increased lipid profile testing, and improved deep vein thrombosis (DVT) prophylaxis.49,50
Comprehensive Stroke Center
The recommendations to establish CSCs were published in 2005.51 In 2011, the ASA published the scientific statement, “Metrics for Measuring Quality of Care in Comprehensive Stroke Centers,” which delineates the set of metrics and related data that CSCs should track to ensure optimal stroke outcome and adherence to current recommendations.10 According to these recommendations, a CSC should be able to offer 24/7 (24 hours per day, 7 days per week) state-of-the-art care on the full spectrum of cerebrovascular diseases. A few states, including New Jersey, Missouri, and Florida, have developed their own legislative efforts to certify PSCs and CSCs. In the fall of 2012, TJC began providing accreditation for CSCs using many of the metrics outlined in the ASA CSC publication.
The data highlighting the patient-centered benefits of integrating CSCs into regional stroke systems of care are emerging. Recently, Orange County, California, organized regional stroke care around CSCs in a hub-and-spoke model, serving just over 3 million people.52 Among patients taken directly to the CSCs in this model, 25.1% received acute reperfusion therapies (intravenous rtPA, endovascular therapies, or both). A recent analysis of 134 441 stroke patients in New Jersey hospitals showed that CSCs had no gap in mortality rate between weekday and weekend admissions, whereas mortality was higher when patients were admitted on weekends at other stroke centers.53 In Finland, where stroke systems of care are organized on a national level, a 7-year study of all stroke patients in the country demonstrated a clear association between the level of acute stroke care and patient outcomes, with the lowest rates of mortality and severe disability seen in CSCs.41
Neurocritical care units are essential elements of CSCs. The need for neurologically focused critical care has expanded rapidly in the past 2 decades in parallel with an increasing understanding of the nature of brain and spinal cord injury, especially the secondary injuries that commonly occur. Improvements in clinical outcome attributable to focused critical care have been documented,54–56 as have a reduction in and an earlier recognition of complications57 and reduced days of hospitalization.54,56 In patients with acute ischemic stroke, admission to neurocritical care units should be considered for those with severe deficits, large-volume infarcts with the potential for significant cerebral edema, significant comorbidities, blood pressure that is difficult to control, or prior intravenous and intra-arterial recanalization interventions.
Acute Stroke-Ready Hospital
ASRHs, previously called stroke-capable hospitals, are hospitals that have made an institutional commitment to effectively and efficiently evaluate, diagnose, and treat most ED stroke patients but that do not have fully organized inpatient stroke systems of care. ASRHs have many of the same elements as a PSC:
Written emergency stroke care protocols
Written transfer agreement with a hospital with neurosurgical expertise
Director of stroke care to oversee hospital stroke policies and procedures (this may be a clinical staff member or the designee of the hospital administrator)
Ability to administer intravenous rtPA
Ability to perform emergency brain imaging (eg, CT scan) at all times
Ability to conduct emergency laboratory testing at all times
Maintenance of a stroke patient log
Additionally, ASRHs have well-developed relationships with regional PSCs and CSCs for additional support. Stroke expertise and neuroimaging interpretation in ASRHs are often in the forms of telemedicine and teleradiology, which require close collaboration within the regional stroke system of care. Many ASRHs do not have sufficient resources to establish and maintain a stroke unit; thus, in some circumstances, once patients are diagnosed and initial treatments delivered, patients are transported to a PSC or CSC. ASRHs are also responsible for EMS stroke education and integration into the stroke system of care. The development of ASRHs has the potential to greatly extend the reach of stroke systems of care into underserved regions.
Telemedicine or “Telestroke”
With the rapid growth of telemedicine for stroke, more data are now available supporting the use of telemedicine to deliver stroke care in regions without local stroke expertise.58,59 Telemedicine (also called telestroke) may help solve the shortage of neurologists and radiologists, allowing hospitals to become acute stroke ready.2,3 Many uses of telemedicine for stroke involve a hub-and-spoke model, in which the hub hospital, often a tertiary stroke center, provides specialty services to spoke hospitals. Telemedicine is integrated audio and visual remote assessment. Telemedicine can provide 24/7 acute stroke expertise to hospitals without full-time neurological or radiological services at the spoke hospital.60 Although the technological sophistication and prices of the systems can vary, it is essential that the system have the capability to provide 2-way real-time audiovisual conferencing and share the images. The benefits of telestroke are several: Telestroke optimizes the use of intravenous rtPA to treat patients in hospitals without an on-site neurologist,61 decreases time to initiate intravenous rtPA, and provides treatment with similar safety as PSCs (symptomatic intracerebral hemorrhage [sICH] in 2%–7%, in-house mortality rate 3.5%).62–65 Although the economic issues regarding the use of telestroke remain to be fully explored, the benefit of telestroke in extending timely stroke care to remote hospitals is clear. These benefits include immediate access to specialty consultations, reliable neurological examinations, and National Institutes of Health Stroke Scale (NIHSS) scores; high rates of intravenous fibrinolysis with low rates of hemorrhage; and mortality rates and functional outcomes of intravenous fibrinolysis comparable to those in randomized trials.66–68 Therefore, when the physical presence of a stroke team physician at the bedside is not possible, telestroke should be established so that additional hospitals can potentially meet the criteria to become ASRHs and PSCs.69,70
Teleradiology is a critical aspect of stroke telemedicine and is defined as the ability to obtain radiographic images at one location and transmit them to another for diagnostic and consultative purposes.71 According to these standards of practice, the Centers for Medicare and Medicaid Services provide reimbursement for both intrastate and interstate teleradiology services,72,73 and the TJC and other accrediting bodies play an important role in the performance, appraisal, and credentialing of teleradiology systems.74 There are only a limited number of studies describing the use of teleradiology to read non–contrast-enhanced CT scans of the brain.75–78 These studies have mainly focused on the feasibility of a teleradiology approach for stroke,79 including some that used personal digital assistants77,78 and smartphones.80,81 One pilot study provided encouraging preliminary evidence that neurologists with stroke expertise can determine radiological intravenous rtPA eligibility via teleradiology.82 Additional studies involving larger samples are necessary to validate these results.
Stroke Care Quality Improvement Process and Establishment of Data Repositories
There is now sufficient literature supporting the initiation of stroke care quality improvement processes. The success of such processes relies on the establishment of quality databases so that data on the performance of quality measurements can be captured. For all certified PSCs, there is an established database to capture the performances on the 8 TJC-mandated quality measures for stroke care. Although all certified PSCs submit their performance data to TJC quarterly, it is beneficial for all hospitals to establish a stroke care data repository. Hospitals can then routinely track their stroke care quality measurements, identify gaps and disparities in providing stroke care, and use these data to design programs to address the gaps or disparities. One such example is the Paul Coverdell National Acute Stroke Registry, which collects data from 8 participating states. Data from the first 4 prototype registries in Georgia, Massachusetts, Michigan, and Ohio showed that overall, 4.51% of ischemic stroke patients were receiving intravenous rtPA on admission.83 By conducting process improvement programs, the Michigan Paul Coverdell National Acute Stroke Registry showed that documentation of the reasons for not giving intravenous rtPA increased by 13%.84 Another example showed that hospitals participating in the Paul Coverdell National Acute Stroke Registry had significant improvements in 9 of the 10 performance measures from 2005 to 2009, with one being that the average annual use of intravenous rtPA increased by 11%.85
Get With The Guidelines (GWTG)-Stroke, provided by the AHA/ASA, is a patient management and data collection tool that ensures continuous quality improvement of acute stroke treatment and stroke prevention. It focuses on care team protocols to ensure that stroke patients are managed according to evidence-based medicine. Currently, there are >1500 hospitals in the United States using the GWTG-Stroke program.86 From 2003 to 2007, a study of 322 847 hospitalized stroke patients in 790 US academic and community hospitals voluntarily participating in the GWTG-Stroke program showed significant improvement in stroke care by participating in the program. Improvements in receipt of guidelines-based care within the 5-year period were as follows: intravenous rtPA use within 2 hours, from 42.9% to 72.84%; antithrombotics within 48 hours of admission, from 91.46% to 97.04%; DVT prophylaxis, from 73.79% to 89.54%; discharged on antithrombotic medication, from 95.68% to 98.88%; anticoagulation for atrial fibrillation, from 95.3% to 98.39%; treatment of low-density lipoprotein cholesterol levels >100 mg/dL, from 73.63% to 88.29%; and smoking cessation efforts with either medication or counseling, from 65.21% to 93.61%.87 A previous study of adherence to evidence-based interventions associated with the process improvement and internet-based data collection showed that the use of intravenous rtPA for patients with ischemic stroke presenting within 2 hours of onset improved from 23.5% to 40.8%. Eleven of 13 quality stroke care measurements showed statistically and clinically significant improvement.88
More recent analysis of the first 1 million patients from 1392 hospitals in GWTG-Stroke showed significant improvements over time from 2003 to 2009 in quality of care (all-or-none measure, 44.0% versus 84.3%; +40.3%, P<0.0001).89 GWTG-Stroke also found disparities in stroke care between men and women. Women received less defect-free care than men (66.3% versus 71.1%; adjusted odds ratio [OR], 0.86; 95% confidence interval [CI], 0.85–0.87) and were less likely to be discharged home (41.0% versus 49.5%; adjusted OR, 0.84; 95% CI, 0.83–0.85).90
Nevertheless, stroke care quality improvement should be an ongoing process for every hospital. One example of this process improvement is to shorten the door-to-needle time to <60 minutes. For every 15-minute reduction of door-to-needle time, there is a 5% lower odds of in-hospital mortality (adjusted OR, 0.95; 95% CI, 0.92–0.98; P=0.0007). However, from this set of GWTG-Stroke data, among 25 504 acute ischemic stroke patients treated with intravenous rtPA within 3 hours of symptom onset at 1082 hospital sites, only 26.6% of patients had a door-to-needle time of the recommended ≤60 minutes.91
Conclusions and Recommendations
All patients with stroke and at risk for stroke benefit from the development of stroke systems of care. States and regions should be encouraged to engage all regional stakeholders to build stroke systems, which in the end will improve patient outcomes through prevention and treatment of stroke, as well as poststroke rehabilitation.
The creation of PSCs is recommended (Class I; Level of Evidence B). The organization of such resources will depend on local resources. The stroke system design of regional ASRHs and PSCs that provide emergency care and that are closely associated with a CSC, which provides more extensive care, has considerable appeal. (Unchanged from the previous guideline13)
Certification of stroke centers by an independent external body, such as TJC or state health department, is recommended (Class I; Level of Evidence B). Additional medical centers should seek such certification. (Revised from the previous guideline13)
Healthcare institutions should organize a multidisciplinary quality improvement committee to review and monitor stroke care quality benchmarks, indicators, evidence-based practices, and outcomes (Class I; Level of Evidence B). The formation of a clinical process improvement team and the establishment of a stroke care data bank are helpful for such quality of care assurances. The data repository can be used to identify the gaps or disparities in quality stroke care. Once the gaps have been identified, specific interventions can be initiated to address these gaps or disparities. (New recommendation)
For patients with suspected stroke, EMS should bypass hospitals that do not have resources to treat stroke and go to the closest facility most capable of treating acute stroke (Class I; Level of Evidence B). (Unchanged from the previous guideline13)
For sites without in-house imaging interpretation expertise, teleradiology systems approved by the Food and Drug Administration (FDA) or equivalent organization are recommended for timely review of brain CT and MRI scans in patients with suspected acute stroke (Class I; Level of Evidence B). (New recommendation)
When implemented within a telestroke network, teleradiology systems approved by the FDA (or equivalent organization) are useful in supporting rapid imaging interpretation in time for fibrinolysis decision making (Class I; Level of Evidence B). (New recommendation)
The development of CSCs is recommended (Class I; Level of Evidence C). (Unchanged from the previous guideline13)
Implementation of telestroke consultation in conjunction with stroke education and training for healthcare providers can be useful in increasing the use of intravenous rtPA at community hospitals without access to adequate onsite stroke expertise (Class IIa; Level of Evidence B). (New recommendation)
The creation of ASRHs can be useful (Class IIa; Level of Evidence C). As with PSCs, the organization of such resources will depend on local resources. The stroke system design of regional ASRHs and PSCs that provide emergency care and that are closely associated with a CSC, which provides more extensive care, has considerable appeal. (New recommendation)
Emergency Evaluation and Diagnosis of Acute Ischemic Stroke
Given the narrow therapeutic windows for treatment of acute ischemic stroke, timely ED evaluation and diagnosis of ischemic stroke are paramount.92,93 Hospitals and EDs should create efficient processes and pathways to manage stroke patients in the ED and inpatient settings. This should include the ability to receive, identify, evaluate, treat, and/or refer patients with suspected stroke, as well as to obtain access to stroke expertise when necessary for diagnostic or treatment purposes.
A consensus panel convened by the National Institutes of Neurological Disorders and Stroke (NINDS) established goals for time frames in the evaluation of stroke patients in the ED.94,95 At this same symposium, the “stroke chain of survival” was promoted as a template for identifying critical events in the ED identification, evaluation, and treatment of stroke patients (Table 5). By using this template and the time goals, hospitals and EDs can create effective systems for optimizing stroke patient care.97
Emergency Triage and Initial Evaluation
ED patients with suspected acute stroke should be triaged with the same priority as patients with acute myocardial infarction or serious trauma, regardless of the severity of neurological deficits. Although specific data on the efficacy of stroke screening tools and scoring systems are lacking for ED triage, the demonstrated utility of such tools in the prehospital environment supports their use in this setting.32,34,98,99 Once in the ED, validated tools for identification of stroke patients within the ED are available.100
The initial evaluation of a potential stroke patient is similar to that of other critically ill patients: immediate stabilization of the airway, breathing, and circulation (ABCs). This is quickly followed by an assessment of neurological deficits and possible comorbidities. The overall goal is not only to identify patients with possible stroke but also to exclude stroke mimics (conditions with strokelike symptoms), identify other conditions that require immediate intervention, and determine potential causes of the stroke for early secondary prevention. Importantly, early implementation of stroke pathways and/or stroke team notification should occur at this point.
The single most important piece of historical information is the time of symptom onset. This is defined as when the patient was at his or her previous baseline or symptom-free state. For patients unable to provide this information or who awaken with stroke symptoms, the time of onset is defined as when the patient was last awake and symptom-free or known to be “normal.”
Establishing onset time may require confirming the patient’s, bystander’s, or EMS personnel’s initial assessment. Creative questioning to establish time anchors potentially allows treatment of patients initially identified as “onset time unknown.” These include inquiring about prestroke or poststroke cellular phone use (and identifying the corresponding call time stamp) or use of television programming times to determine onset time. Patients with “wake-up” strokes may identify a time point when they were ambulatory to the bathroom or kitchen.
Often a patient’s current symptoms were preceded by similar symptoms that subsequently resolved. For patients who had neurological symptoms that completely resolved, the therapeutic clock is reset, and the time of symptom onset begins anew. However, the longer the transient neurological deficits last, the greater the chance of detecting neuroanatomically relevant focal abnormalities on diffusion-weighted and apparent diffusion coefficient imaging.75 Whether this represents an increased risk of hemorrhage with fibrinolysis remains to be determined.
Additional historical items include circumstances surrounding the development of the neurological symptoms and features that may point to other potential causes of the symptoms. Although not absolutely accurate, some early historical data and clinical findings may direct the physician toward an alternate diagnosis of another cause for the patient’s symptoms (Table 6). It is important to ask about risk factors for arteriosclerosis and cardiac disease, as well as any history of drug abuse, migraine, seizure, infection, trauma, or pregnancy. Historical data related to eligibility for therapeutic interventions in acute ischemic stroke are equally important. Bystanders or family witnesses should be asked for information about onset time and historical issues as well, and EMS personnel should be encouraged to identify witnesses and bring them with the patient. This is of particular importance when patients are unable to provide a history.
After the airway, breathing, and circulation have been assessed and specific vital signs determined, such as blood pressure, heart rate, oxygen saturation, and temperature, a more deliberate and detailed physical examination is performed. The detailed physical examination may be conducted by the emergency physician, the stroke expert, or both. The general examination is important to identify other potential causes of the patients’ symptoms, potential causes of an ischemic stroke, coexisting comorbidities, or issues that may impact the management of an ischemic stroke. Examination of the head and face may reveal signs of trauma or seizure activity. Auscultation of the neck may reveal carotid bruits; palpation, auscultation, and observation may reveal signs of congestive heart failure. Auscultation of the chest similarly may reveal cardiac murmurs, arrhythmias, and rales. A general examination of the skin may reveal stigmata of coagulopathies, platelet disorders, signs of trauma, or embolic lesions (Janeway lesions, Osler nodes). A thorough examination to identify acute comorbidities and conditions that may impact treatment selection is important.
Neurological Examination and Stroke Scale/Scores
The initial neurological examination should be brief but thorough. At this point, if the initial history and brief examination are suggestive of a stroke, stroke code activation should occur. The use of a standardized neurological examination ensures that the major components of a neurological examination are performed in a timely and uniform fashion. Formal stroke scores or scales, such as the NIHSS or Canadian Neurological Scale, may be performed rapidly, have demonstrated utility, and may be administered by a broad spectrum of healthcare providers (Table 7).101,102 Use of a standardized assessment and stroke scale helps quantify the degree of neurological deficits, facilitate communication, identify the location of vessel occlusion, provide early prognosis, help select patients for various interventions, and identify the potential for complications.103–105
Although strokes are the most common cause of new focal neurological deficits, other causes must be considered as well in the acute setting. Stroke mimics were identified in ≈3% of patients in 2 series of patients treated with fibrinolytics, with seizures and conversion disorder identified most frequently.106,107 No evidence of increased fibrinolytic treatment risk, however, was identified for these patients. More recently, Chernyshev et al108 reported from their registry of 512 patients treated with intravenous rtPA for presumed ischemic stroke within 3 hours from symptom onset that 21% were later determined to be stroke mimics. In this cohort composed largely of patients with seizures, complicated migraines, and conversion disorders, none experienced a symptomatic hemorrhage, and 87% were functionally independent at discharge. Important conditions mimicking stroke and their clinical features are listed in Table 6. Despite the lack of apparent harm of intravenous rtPA in stroke mimics, an accompanying editorial suggested stroke mimic treatment rates at experienced centers should be <3% using noncontrast CT alone.109 Means for striking a balance between speed to treatment and diagnostic accuracy will continue to evolve.
Access to Neurological Expertise
Patients in many hospital settings have limited access to specialists with stroke expertise. Although evidence supporting the utility of acute “code stroke” teams and telestroke systems is plentiful, their availability is dependent on local resources. The evidence on the safety of fibrinolytic delivery without a neurologist stroke specialist present in person or by telemedicine is less robust.
Although emergency physicians exhibit high sensitivity and positive predictive value in identifying patients with stroke,110,111 only 6 studies112–117 have identified instances of fibrinolytic delivery in the setting of acute stroke by an emergency or primary care physician (either alone or in telephone consultation with a neurologist). The number of patients treated by nonneurologists in these studies was small, ranging from 6 to 53. Two additional studies reported cautionary findings for “community models” of acute stroke care, in which care is delivered outside an acute stroke team. One study noted an increase in sICH in a series of 70 patients treated by community neurologists,118 and both found increased in-hospital mortality among intravenous rtPA–treated stroke patients.118,119 In the case of the Cleveland, OH, experience, these poor outcomes led to quality improvement initiatives that decreased overall rates of symptomatic hemorrhage from 15.7% to 6.4%.120
Larger, more recent studies, however, found no evidence of increased risk for mortality, intracerebral hemorrhage (ICH), or reduced functional recovery with a variety of acute response arrangements in a US series of 273 consecutive stroke patients treated with fibrinolytics. These patients were treated by 95 emergency physicians from 4 hospitals without an acute fibrinolytic stroke team over a 9-year period.121 One third of the cases were treated without a neurological consultation, with a telephone consultation only, or with an in-person consultation, respectively. An ongoing National Institutes of Health–supported study (Increasing Stroke Treatment Through Interventional Behavior Change Tactics [INSTINCT]) is expected to accrue >500 intravenous rtPA–treated patients in a randomly selected cohort of 24 Michigan hospitals and will provide a comprehensive assessment of the safety of intravenous rtPA use in the community ED setting.122
Thus, current data support multiple approaches to obtaining specialist consultation when needed in the setting of acute stroke. These range from using committed local physicians to using telephones and telemedicine (integrated audio and visual remote assessment) to access local or regional specialists or activating an acute stroke team. Development of local stroke processes to maximize available local and regional resources and to clearly identify access to neurological expertise optimizes opportunities for acute treatment.
Several tests should be routinely emergently performed as indicated in patients with suspected ischemic stroke, primarily to exclude important alternative diagnoses (especially ICH), assess for serious comorbid diseases, aid in treatment selection, and search for acute medical or neurological complications of stroke (Table 8). Laboratory tests to consider in all patients include blood glucose, electrolytes with renal function studies, complete blood count with platelet count, cardiac markers, prothrombin time (PT), international normalized ratio (INR), and activated partial thromboplastin time (aPTT). Hypoglycemia may cause focal signs and symptoms that mimic stroke, and hyperglycemia is associated with unfavorable outcomes. Determination of the platelet count and, in patients taking warfarin or with liver dysfunction, the PT/INR is important. Cardiac markers are frequently elevated in acute ischemic stroke, with elevations occurring in 5% to 34% of patients, and these elevations have prognostic significance.123 Elevation of cardiac troponin T is associated with increased stroke severity and mortality risk, as well as worse clinical outcomes.124–127
Certain laboratory tests should be considered in select patients. As the use of direct thrombin inhibitors, such as dabigatran, and direct factor Xa inhibitors, such as rivaroxaban and apixaban, becomes more prevalent, it is important to understand what studies may assist in determining qualitatively whether an anticoagulant effect is present. The PT/INR is not helpful in determining whether an anticoagulant effect from dabigatran is present. A patient may have significant concentrations without alterations in PT/INR. A thrombin time (TT) is a sensitive indicator to the presence of dabigatran activity, and a normal TT excludes the presence of significant activity; however, it may be influenced by the use of other anticoagulants. The ecarin clotting time (ECT) demonstrates a linear relationship with direct thrombin inhibitor levels, and a normal ECT generally excludes a significant direct thrombin inhibitor effect and is not influenced by other anticoagulants; however, this test may not be available at all hospitals.128 As newer anticoagulation agents become available, for instance, direct factor Xa inhibitors, specific assays of activity may be required.
Beyond new anticoagulants, specific laboratory tests may be helpful when there is a suspicion of drug abuse, particularly in cases of stroke in young adults. In this instance, toxicological screens for sympathomimetic use (cocaine, methamphetamine, etc) may identify the underlying cause of the stroke.129 Although uncommon, women of childbearing age with acute stroke may be pregnant, and results from pregnancy testing may impact the patient’s overall management. Examination of the cerebrospinal fluid has a limited role in the acute evaluation of patients with suspected stroke, unless there is a strong suspicion for subarachnoid hemorrhage or acute central nervous system infections.
Because time is critical, fibrinolytic therapy should not be delayed while awaiting the results of the PT, aPTT, or platelet count unless a bleeding abnormality or thrombocytopenia is suspected, the patient has been taking warfarin and heparin, or anticoagulation use is uncertain. Retrospective reviews of patients who received intravenous fibrinolysis demonstrated very low rates of unsuspected coagulopathies and thrombocytopenia that would have constituted a contraindication to fibrinolysis.130,131 The only laboratory result required in all patients before fibrinolytic therapy is initiated is a glucose determination; use of finger-stick measurement devices is acceptable.
Chest radiography is often performed in patients with acute stroke; however, only limited observational data are available to guide decision making regarding its utility. One study that evaluated chest radiographs obtained 12 to 24 hours after admission for stroke found clinical management was altered in 3.8% of cases.132 A different study found 3.8% of routine chest radiographs obtained during a code stroke activation (within 6 hours of symptom onset) had a potentially relevant abnormality, with 1 film showing a possibly wide mediastinum (subsequently determined to be normal) and 1.8% having confirmed pulmonary opacities. Thus, the utility of routine chest radiography is debatable in the absence of clinical suspicion of underlying pulmonary, cardiac, or vascular disease.133 As with diagnostic laboratory tests, chest radiography should not delay administration of intravenous rtPA unless there are specific concerns about intrathoracic issues, such as aortic dissection.
All acute stroke patients should undergo cardiovascular evaluation, both for determination of the cause of the stroke and to optimize immediate and long-term management. This cardiac assessment should not delay reperfusion strategies. Atrial fibrillation may be seen on an admission electrocardiogram; however, its absence does not exclude the possibility of atrial fibrillation as the cause of the event. Thus, ongoing monitoring of cardiac rhythm on telemetry or by Holter monitoring may detect atrial fibrillation or other serious arrhythmias.134,135 Acute stroke and acute myocardial infarction can present contemporaneously, with one precipitating the other. Ischemic stroke can also cause electrocardiogram abnormalities and, occasionally, cardiac decompensation (cardiomyopathy) via neurohormonal pathways.136–139
Because of the close association between stroke and cardiac abnormalities, it is important to assess the cardiovascular status of patients presenting with acute stroke. Baseline electrocardiogram and cardiac biomarkers may identify concurrent myocardial ischemia or cardiac arrhythmias. Troponin is preferred because of its increased sensitivity and specificity over creatine phosphokinase or creatine phosphokinase–MB. Repeat electrocardiogram and serial cardiac enzymes may identify developing silent ischemia or paroxysmal arrhythmias not detected on initial studies.
Conclusions and Recommendations
The evaluation and initial treatment of patients with stroke should be performed expeditiously. Organized protocols and the availability of a stroke team speed the clinical assessment, the performance of diagnostic studies, and decisions for early management. The clinical assessment (history, general examination, and neurological examination) remains the cornerstone of the evaluation. Stroke scales, such as the NIHSS, provide important information about the severity of stroke and prognostic information and influence decisions about acute treatment.
Because time is critical, a limited number of essential diagnostic tests are recommended. Additional diagnostic studies, including cardiac and vascular imaging, often are time consuming and may delay emergency treatment. Stroke protocols and pathways should clearly define which tests must be performed before acute treatment decisions and which may be performed subsequent to acute stroke therapies.
An organized protocol for the emergency evaluation of patients with suspected stroke is recommended (Class I; Level of Evidence B). The goal is to complete an evaluation and to begin fibrinolytic treatment within 60 minutes of the patient’s arrival in an ED. Designation of an acute stroke team that includes physicians, nurses, and laboratory/radiology personnel is encouraged. Patients with stroke should have a careful clinical assessment, including neurological examination. (Unchanged from the previous guideline)
The use of a stroke rating scale, preferably the NIHSS, is recommended (Class I; Level of Evidence B). (Unchanged from the previous guideline13)
A limited number of hematologic, coagulation, and biochemistry tests are recommended during the initial emergency evaluation, and only the assessment of blood glucose must precede the initiation of intravenous rtPA (Table 8) (Class I; Level of Evidence B). (Revised from the previous guideline13)
Baseline electrocardiogram assessment is recommended in patients presenting with acute ischemic stroke but should not delay initiation of intravenous rtPA (Class I; Level of Evidence B). (Revised from the previous guideline13)
Baseline troponin assessment is recommended in patients presenting with acute ischemic stroke but should not delay initiation of intravenous rtPA (Class I; Level of Evidence C). (Revised from the previous guideline13)
The usefulness of chest radiographs in the hyperacute stroke setting in the absence of evidence of acute pulmonary, cardiac, or pulmonary vascular disease is unclear. If obtained, they should not unnecessarily delay administration of fibrinolysis (Class IIb; Level of Evidence B). (Revised from the previous guideline13)
Early Diagnosis: Brain and Vascular Imaging
Timely brain imaging and interpretation remains critical to the rapid evaluation and diagnosis of patients with potential ischemic strokes. Newer strategies are playing an increasingly important role in the initial evaluation of patients with acute stroke. Brain imaging findings, including the size, location, and vascular distribution of the infarction, the presence of bleeding, severity of ischemic stroke, and/or presence of large-vessel occlusion, affect immediate and long-term treatment decisions. Information about the possible degree of reversibility of ischemic injury, intracranial vessel status (including the location and size of occlusion), and cerebral hemodynamic status can be obtained by modern imaging studies.140,141 Although these modalities are increasingly available emergently, non–contrast-enhanced computed tomography (NECT) remains sufficient for identification of contraindications to fibrinolysis and allows patients with ischemic stroke to receive timely intravenous fibrinolytic therapy. NECT should be obtained within 25 minutes of the patient’s arrival in the ED.
Parenchymal Brain Imaging
NECT and Contrast-Enhanced CT Scans of the Brain
NECT definitively excludes parenchymal hemorrhage and can assess other exclusion criteria for intravenous rtPA, such as widespread hypoattenuation.142–145 NECT scanning of the brain accurately identifies most cases of intracranial hemorrhage and helps discriminate nonvascular causes of neurological symptoms (eg, brain tumor). NECT may demonstrate subtle visible parenchymal damage within 3 hours.146–148 NECT is relatively insensitive in detecting acute and small cortical or subcortical infarctions, especially in the posterior fossa.75 Despite these limitations, its widespread immediate availability, relative ease of interpretation, and acquisition speed make NECT the most common modality used in acute ischemic stroke imaging.
With the advent of intravenous rtPA treatment, interest has grown in using NECT to identify subtle, early signs of ischemic brain injury (early infarct signs) or arterial occlusion (hyperdense vessel sign) that might affect decisions about treatment. A sign of cerebral ischemia within the first few hours after symptom onset on NECT is loss of gray-white differentiation.76–78,149,150 This sign may manifest as loss of distinction among the nuclei of the basal ganglia (lenticular obscuration) or as a blending of the densities of the cortex and underlying white matter in the insula (insular ribbon sign)150 and over the convexities (cortical ribbon sign). Another sign of cerebral ischemia is swelling of the gyri that produces sulcal effacement. The more rapidly these signs become evident, the more profound the degree of ischemia. However, the ability of observers to detect these early infarct signs on NECT is quite variable and occurs in ≤67% of cases imaged within 3 hours. Detection is influenced by the size of the infarct, severity of ischemia, and the time between symptom onset and imaging.151,152 Detection may increase with the use of a structured scoring system such as the Alberta Stroke Program Early CT Score (ASPECTS) or the CT Summit Criteria,151–155 as well as with the use of better CT “windowing and leveling” to differentiate between normal and abnormal tissues.156
Another useful CT sign is that of increased density within the occluded artery, such as the hyperdense middle cerebral artery (MCA) sign, indicative of large-vessel occlusion.157 Large-vessel occlusion typically causes severe stroke, independently predicts poor neurological outcome,157–159 and is a stronger predictor of “neurological deterioration” (91% positive predictive value) than even early CT evidence of >50% MCA involvement (75% positive predictive value).159,160 The hyperdense MCA sign, however, is seen in only one third to one half of cases of angiographically proven thromboses160,161; hence, it is an appropriate indicator of thrombus when present. Another NECT sign is the hyperdense MCA “dot” sign.162 The MCA dot sign represents a clot within a branch of the MCA and is thus typically smaller than the thrombus volume in the MCA and possibly a better target for intravenous rtPA. Barber et al162 found that patients with the MCA dot sign alone had better outcomes than patients with a hyperdense MCA sign. Validation for the MCA dot sign has been performed with angiography, with the conclusion that the sensitivity is low (38%) but the specificity is 100%.163 The hyperdense basilar artery sign has been described with similar implications as the hyperdense MCA sign.164,165
The presence, clarity, and extent of early ischemia and infarction on NECT are correlated with a higher risk of hemorrhagic transformation after treatment with fibrinolytic agents. In combined data from 2 trials of intravenous rtPA administered within 3 hours of symptom onset, NECT evidence of early clear hypodensity or mass effect was accompanied by an 8-fold increase in the risk of symptomatic hemorrhage.166 In a second analysis, more subtle early infarct signs involving more than one third of the territory of the MCA were not independently associated with increased risk of adverse outcome after intravenous rtPA treatment, and as a group, these patients still benefited from therapy.148 In a European trial in which fibrinolytic therapy was administered within 6 hours of symptom onset, patients estimated to have involvement of more than one third of the territory of the MCA had an increased risk of ICH, whereas those with less involvement benefited the most from fibrinolytic treatment.144,167 Because of this increased hemorrhage risk, patients with involvement of more than one third of the territory of the MCA by early ischemic signs were excluded from entry in the pivotal trial confirming the benefit of intravenous fibrinolytic therapy in the 3- to 4.5-hour window and the major trials of intra-arterial fibrinolysis up to 6 hours after onset.168–170
MRI of the Brain
Standard MRI sequences (T1 weighted, T2 weighted, fluid-attenuated inversion recovery [FLAIR]) are relatively insensitive to the changes of acute ischemia.171 Diffusion-weighted imaging (DWI) has emerged as the most sensitive and specific imaging technique for acute infarct, far better than NECT or any other MRI sequence. DWI has a high sensitivity (88% to 100%) and specificity (95% to 100%) for detecting infarcted regions, even at very early time points,172–174 within minutes of symptom onset.172,175–181 DWI allows identification of the lesion size, site, and age. DWI can detect relatively small cortical lesions and small deep or subcortical lesions, including those in the brain stem or cerebellum, areas often poorly or not visualized with standard MRI sequences and NECT scan techniques.182–185 DWI can identify subclinical satellite ischemic lesions that provide information on stroke mechanism.173,176,179,186–197 There are a few articles describing negative DWI studies when cerebral perfusion is decreased enough to produce infarction198,199 and the reversal, partial or complete, of DWI abnormalities with restoration of perfusion.200 Thus, early after ischemia onset, the visible diffusion lesion will include both regions of irreversible infarction with more severe apparent diffusion coefficient changes and regions of salvageable penumbra with less severe apparent diffusion coefficient changes.
The artery susceptibility sign is the magnetic resonance (MR) correlate of the hyperdense MCA seen on NECT. A direct comparison of NECT and MRI in patients with occlusion of the proximal MCA found that 54% of patients demonstrated this sign on NECT, whereas 82% of the same patients had clot demonstrated on MRI using a gradient echo sequence.161 Vascular hyperintensities on fluid-attenuated inversion recovery sequences can indicate slow-flowing blood passing through leptomeningeal collaterals.201 Conventional MRI is more sensitive than standard NECT in identifying both new and preexisting ischemic lesions in patients with 24-hour time-defined TIAs.202–220 Multiple series show convergent results regarding the frequency of DWI positivity among time-defined TIA patients; among 19 studies that included 1117 patients with TIA, the aggregate rate of DWI positivity was 39%, with frequency by site ranging from 25% to 67%. DWI-positive lesions tend to be smaller and multiple in TIA patients.75 There does not appear to be a predilection for cortical or subcortical regions or particular vascular territories. Recently, several studies have demonstrated that DWI positivity in TIA patients is associated with a higher risk of recurrent ischemic events.221–223
The appearance of hemorrhage on MRI is dependent on both the age of the blood and the pulsing sequences used.224–231 Magnetic susceptibility imaging is based on the ability of a T2*-weighted MR sequence to detect very small amounts of deoxyhemoglobin, in addition to other compounds such as those containing iron or calcium. Two prospective studies demonstrated that MRI was as accurate as NECT in detecting hyperacute intraparenchymal hemorrhage in patients presenting with stroke symptoms within 6 hours of onset when gradient echo sequences were used.228,232 Accordingly, MRI may be used as the sole initial imaging modality to evaluate acute stroke patients, including candidates for fibrinolytic treatment. Gradient echo sequences also have the ability to detect clinically silent prior microbleeds not visualized on NECT. Some data suggest that microbleeds represent markers of bleeding-prone angiopathy and increased risk of hemorrhagic transformation after antithrombotic and fibrinolytic therapy.233–235 However, other studies have not found an increased risk in patients with small numbers of microbleeds.236 The importance of the presence of large numbers of microbleeds on MRI in fibrinolytic decision making remains uncertain.
Compared with CT, advantages of MRI for parenchymal imaging include the ability to distinguish acute, small cortical, small deep, and posterior fossa infarcts; the ability to distinguish acute from chronic ischemia; identification of subclinical satellite ischemic lesions that provide information on stroke mechanism; the avoidance of exposure to ionizing radiation; and greater spatial resolution. Limitations of MRI in the acute setting include cost, relatively limited availability of the test, relatively long duration of the test, increased vulnerability to motion artifact, and patient contraindications such as claustrophobia, cardiac pacemakers, patient confusion, or metal implants. Additionally, in ≈10% of patients, an inability to remain motionless may obviate the ability to obtain a quality MRI.
Intracranial Vascular Imaging
An important aspect of the workup of patients with stroke, TIA, or suspected cerebrovascular disease is imaging of intracranial vasculature. The majority of large strokes are caused by occlusion in ≥1 large vessel. Large-vessel occlusion is a devastating condition.158,159,237–249 Detection of large-vessel occlusion by means of noninvasive intracranial vascular imaging greatly improves the ability to make appropriate clinical decisions.168,170,237,239,241,250 It is also essential to establish as soon as possible the mechanism of ischemia to prevent subsequent episodes. Large-vessel occlusion can be identified by NECT as described above (hyperdense MCA sign, etc). The length of a clot within the MCA has been directly related to the success of recanalization with intravenous rtPA.251
Helical CT angiography (CTA) provides a means to rapidly and noninvasively evaluate the intracranial and extracranial vasculature in acute, subacute, and chronic stroke settings and thus to provide potentially important information about the presence of vessel occlusions or stenoses.242,252 The accuracy of CTA for evaluation of large-vessel intracranial stenoses and occlusions is very high,253–256 and in some cases its overall accuracy approaches or exceeds that of digital subtraction angiography (DSA).253,257 The sensitivity and specificity of CTA for the detection of intracranial occlusions ranges between 92% and 100% and between 82% and 100%, respectively, with a positive predictive value of 91% to 100%.242,258–260 Because CTA provides a static image of vascular anatomy, it is inferior to DSA for the demonstration of flow rates and direction.
Direct comparisons of CTA source images (CTA-SI) and MRI/DWI have demonstrated very similar sensitivity of these 2 techniques for detecting ischemic regions, with DWI being better at demonstrating smaller abnormalities (reversible or irreversible) and those in the brainstem and posterior fossa.261,262 In one study, CTA-SI was superior in stroke identification for readers with all levels of experience.263 Improved stroke detection explains the greater predictive value for final infarct size by use of CTA-SI.248 For early strokes (<3 hours), CTA-SI ASPECTS has a greater sensitivity to ischemic changes and more accurately identifies the volume of tissue that will ultimately become infarcted than NECT alone.159,248 CTA-SI is more an estimate of cerebral blood volume than the expression of cytotoxic edema seen on NECT.
Intracranial MR angiography (MRA) is performed in combination with brain MRI in the setting of acute stroke to guide therapeutic decision making.264 There are several different MRA techniques that are used for imaging intracranial vessels. They include 2-dimensional time of flight (TOF), 3-dimensional TOF, multiple overlapping thin-slab acquisition, and contrast-enhanced MRA.265 Intracranial MRA with nonenhanced TOF techniques has a sensitivity ranging from 60% to 85% for stenoses and from 80% to 90% for occlusions compared with CTA or DSA.253,258 Typically, TOF MRA is useful in identifying acute proximal large-vessel occlusions but cannot reliably identify distal or branch occlusions.266
Transcranial Doppler (TCD) ultrasonography has been used to detect intracranial vessel abnormalities.267,268 TCD has been used to evaluate occlusions and stenoses in intracranial vessels. TCD accuracy is less than that of CTA and MRA for steno-occlusive disease, with a sensitivity and specificity of TCD ranging from 55% to 90% and from 90% to 95%, respectively.269–276 TCD can detect microembolic signals, which are seen with extracranial or cardiac sources of embolism.277–279
In an attempt to better define the accuracy rate of TCD for intracranial stenoses (a common cause of stroke), the Stroke Outcomes and Neuroimaging of Intracranial Atherosclerosis (SONIA) Trial was designed to evaluate the controlled patient population in the Warfarin-Aspirin Symptomatic Intracranial Disease Study (WASID).276 SONIA enrolled 407 patients at 46 sites. For 50% to 99% stenoses that were angiographically confirmed (the “gold standard”), TCD was able to positively predict 55% of these lesions but was able to rule out 83% of vessels that had <80% stenosis (a low hurdle). This multi-institutional study suggested less than optimal TCD accuracy.276 TCD is more accurate for proximal M1 than distal M1 or M2 disease.256
TCD has been shown to predict, as well as enhance, intravenous rtPA outcomes.280 Large-vessel occlusions and more proximal occlusions identified by TCD have been predictive of poor revascularization results with intravenous rtPA and worse clinical outcomes.281,282 In the presence of an appropriate bone window and for vessels capable of visualization by sonography, TCD has been used to monitor the response of cerebral vessels to fibrinolytic therapy over time, as well as to augment such therapy using ultrasonic energy to enhance clot lysis280,283–286; TCD provides continuous, real-time imaging and can thus determine the timing of recanalization and the occurrence of reocclusion of vessels capable of visualization by sonography.282,284,285,287–290 CLOTBUST (Combined Lysis of Thrombus in Brain Ischemia Using Transcranial Ultrasound and Systemic rtPA) indicated recanalization improvement with continuous TCD but was underpowered to detect a significant final clinical improvement. Although higher-frequency ultrasound appeared safe as a lytic enhancer in CLOTBUST, the Transcranial Low-Frequency Ultrasound-Mediated Thrombolysis in Brain Ischemia study (TRUMBI)291 indicated an increased risk of hemorrhage with low-frequency ultrasound. However, the usefulness of TCD is limited in patients with poor bony windows, and its overall accuracy is dependent on the experience of the technician, the interpreter, and the patient’s vascular anatomy. For posterior circulation stroke, Doppler ultrasound is not helpful; CTA, MRA, or a conventional angiogram is required.
DSA remains the “gold standard” for the detection of many types of cerebrovascular lesions and diseases.292–294 For most types of cerebrovascular disease, the resolution, sensitivity, and specificity of DSA equal or exceed those of the noninvasive techniques, including for arterial stenoses.292,294–298 However, if noninvasive imaging provides firm diagnostic findings, cerebral angiography may not be required.
DSA is an invasive test and can cause serious complications such as stroke and death, although recent advances in high-resolution rapid-sequence digital subtraction imaging, digital image reconstruction with 3-dimensional techniques, catheter technology, and nonionic contrast media have made cervicocerebral angiography easier and safer over the past 2 decades. Most large series have reported rates of stroke or death in <1% of DSA procedures.299–301 The largest series of cases to date reported a rate of stroke or death of <0.2%.299–301 Cerebral angiography need not be the initial imaging modality for emergency intracerebral evaluation of large-vessel occlusion in stroke because of the time necessary to perform the examination; a CTA or MRA can be performed in an additional 2 to 4 minutes during initial stroke evaluation (in a multimodal evaluation in process) and can obviate the need for catheter angiography.292,294
Extracranial Vascular Imaging
It is important to evaluate the extracranial vasculature after the onset of acute cerebral ischemia (stroke or TIA) to aid in the determination of the mechanism of the stroke and thus potentially to prevent a recurrence.6,302 In addition, carotid endarterectomy (CEA) or angioplasty/stenting is occasionally performed acutely, which requires appropriate imaging. The major extracranial cerebral vessels can be imaged by several noninvasive techniques, such as ultrasound, CTA, TOF and contrast-enhanced MRA, and DSA.303–305 Although each technique has certain advantages in specific clinical situations, the noninvasive techniques show general agreement to DSA in 85% to 90% of cases. For evaluation of the degree of stenosis and for determination of patient eligibility for CEA or carotid angioplasty and stenting, DSA is the “gold standard” imaging modality. The use of 2 concordant noninvasive techniques (among ultrasound, CTA, and MRA) to assess treatment candidacy has the advantage of avoiding catheterization risks.306,307 CTA (in the absence of heavy calcifications) and multimodal MRI (including MRA and fat-saturation axial T1 imaging) are highly accurate for detecting dissection; for subtle dissections, DSA and multimodal MRI are complementary, and there have been reports of dissections detected by one modality but not the other.308,309 A very high-grade stenosis (“string sign”) is most accurately detected by DSA, followed closely by CTA and contrast-enhanced MRA.310
Carotid Doppler Ultrasound
Carotid ultrasound is a safe and inexpensive screening technique for imaging the carotid bifurcation and measuring blood velocities.303,311,312 Doppler measures that have been correlated with angiographic stenosis include internal carotid artery peak systolic velocity and end-diastolic velocity, as well as ratios of internal carotid artery and common carotid artery peak systolic velocity.313 Doppler test results and diagnostic criteria are influenced by several factors, such as the equipment, the specific laboratory, and the technologist performing the test.314,315 For these reasons, it is recommended that each laboratory validate its own Doppler criteria for clinically relevant stenosis.316,317 Sensitivity and specificity of carotid ultrasound for detecting lesions >70% are less than for other modalities, in the range of 83% to 86% for sensitivity and 87% to 99% for specificity.318–320 Carotid ultrasound has limited ability to image the extracranial vasculature proximal or distal to the bifurcation.
CTA is a sensitive, specific, and accurate technique for imaging the extracranial vasculature. CTA is clearly superior to carotid ultrasound for differentiating a carotid occlusion from a very high-grade stenosis321 and has been reported to have an excellent (100%) negative predictive value for excluding >70% stenosis compared with catheter angiography, thereby functioning well as a screening test.322 A large meta-analysis found it to have a sensitivity >90% and specificity >95% for detecting significant lesions compared with DSA.255,319,323–326
Two-dimensional and 3-dimensional TOF MRA used for the detection of extracranial carotid disease (threshold stenosis typically 70%) showed a mean sensitivity of 93% and a mean specificity of 88%.265 Contrast-enhanced MRA is more accurate than nonenhanced TOF techniques, with specificities and sensitivities of 86% to 97% and 62% to 91%, respectively, compared with DSA.320,327–332 Craniocervical arterial dissections of the carotid and vertebral arteries can often be detected with MRA.333–336 Contrast-enhanced MRA may improve the detection of arterial dissections,337 although there are few large, prospective studies to prove its accuracy versus catheter angiography. Nonenhanced T1-weighted MRI with fat-saturation techniques can frequently depict a subacute hematoma within the wall of an artery, which is highly suggestive of a recent dissection.338,339 However, an acute intramural hematoma may not be well visualized on fat-saturated T1-weighted MRI until the blood is metabolized to methemoglobin, which may require a few days after ictus. MRA is also helpful for detecting other less common causes of ischemic stroke or TIAs such as arterial dissection, fibromuscular dysplasia, venous thrombosis, and some cases of vasculitis.337
DSA remains the most informative technique for imaging the cervical carotid and vertebral arteries, particularly when making decisions about invasive therapies. In addition to providing specific information about a vascular lesion, DSA can provide valuable information about collateral flow, perfusion status, and other occult vascular lesions that may affect patient management.292–298 As mentioned above, DSA is associated with a risk, albeit small (<1%), of serious complications such as stroke or death.299–301 Catheter angiography can be particularly useful in cases of carotid dissection, both to image the dissection and to delineate the collateral supply to the brain.
Perfusion CT and MRI
In recent years, it has become apparent that information about the nature and severity of the ischemic insult may be just as important as the “time” of the ischemic event for predicting outcome and making therapeutic judgments. There is a growing body of literature positing that ischemic, potentially salvageable “penumbral” tissue is an ideal target for reperfusion and neuroprotective strategies but requires proper patient selection.159,247,262,282,340–344 However, in the acute stroke setting, there is a trade-off between the increased information provided by perfusion imaging and the increased time needed to acquire additional imaging sequences. The performance of these additional imaging sequences should not unduly delay treatment with intravenous rtPA in the ≤4.5-hour window in appropriate patients.283,286,292,297–301
Brain perfusion imaging provides information about regional cerebral hemodynamics in the form of such parameters as cerebral blood flow, cerebral blood volume, and mean transit time. Perfusion CT and perfusion-weighted MRI have been widely incorporated into acute multimodal imaging protocols. Combined with parenchymal imaging, perfusion-weighted MRI or perfusion CT imaging permits delineation of the ischemic penumbra.213,215,216,218,345–349 Perfusion imaging can also indicate areas that are severely and probably irretrievably infarcted. A current technical challenge is that methods for processing of perfusion data to derive perfusion parameters vary, and the most biologically salient perfusion parameters and thresholds for acute decision making have not been fully defined.218 On MRI, the ischemic penumbra is roughly indexed as the area of perfusion-weighted imaging–DWI mismatch.176,203,205,214 On perfusion CT imaging, the penumbra is indexed as the area of mean transit time–cerebral blood volume mismatch.202,210,212,219 “Core” ischemia can be defined accurately by perfusion CT depending on equipment and programming. Various studies have used different hemodynamic parameters, such as mean transit time, cerebral blood volume, and cerebral blood flow,252–258,260,264–275, 350 different thresholds for determining hemodynamic abnormality (eg, degree of reduction in cerebral blood volume and absolute versus relative threshold), and different thresholds for the amount of penumbral tissue that warrants treatment (eg, 20%, 100%, or 200% the size of the infarct core).206,207,213,215–217,347–349 The International Stroke Imaging Repository (STIR) consortium is currently addressing these issues and is attempting to standardize imaging methodology, processing, and interpretation.218
Advantages of the multimodal CT approach over MRI include wider availability of emergency CT imaging, more rapid imaging, and fewer contraindications to CT versus MRI.351–353 Perfusion CT parameters of cerebral blood volume, cerebral blood flow, and mean transit time can be more easily quantified than their perfusion-weighted MRI counterparts, owing in part to the linear relationship between iodinated CT contrast concentration and resulting CT image density, a relationship that does not hold for gadolinium concentration versus MRI signal intensity. Because of its availability and greater degree of quantification, perfusion CT has the potential to increase patient access to new treatments and imaging-based clinical trials.
Disadvantages of the CT approach over MRI include the use of ionizing radiation and iodinated contrast, which carries a small risk of nephrotoxicity. Use of low-osmolar or iso-osmolar contrast minimizes the risk of contrast-induced nephropathy.354,355 A recent study of CTA in patients with acute ischemic and hemorrhagic stroke demonstrated a very low rate of contrast-induced nephropathy (3%), and no patients required dialysis.356 Another disadvantage of perfusion CT is limited brain coverage, typically a 4-cm-thick slab per contrast bolus.242,259,357,358 Developments such as the toggling-table technique allow doubling of the perfusion CT coverage (typically up to 8 cm).359 Finally, the latest generations of the 256- and 320-slice CT scanners afford whole-brain coverage but are limited in availability.
The major advantages of perfusion MRI over perfusion CT include its inclusion in a package of imaging sequences that effectively evaluate many aspects of the parenchyma, including the presence of infarction with DWI, and the avoidance of ionizing radiation. Of note, the whole-brain coverage offered by perfusion MRI comes at the cost of a limited spatial resolution (matrix size or interslice gap) or temporal resolution. Disadvantages of perfusion MRI include limited availability in emergency settings, duration of the study, and patient contraindications such as claustrophobia, cardiac pacemakers, patient confusion and/or motion, or metal implants. Gadolinium reactions are uncommon but can be dangerous.353,360 Nephrogenic systemic fibrosis/nephrogenic fibrosing dermopathy is caused by gadolinium-based contrast agents used for MRI.360 Gadolinium-based MR contrast media generally should be avoided in the presence of advanced renal failure with estimated glomerular filtration rate <30 mL·min−1.73·m−2.360,361 Arterial spin labeling is an MRI method that assesses brain perfusion without the need to inject gadolinium contrast material, but it is not widely available.277
Several recent trials have studied MRI perfusion/diffusion mismatch. EPITHET (Echoplanar Imaging Thrombolytic Evaluation Trial) was designed to answer the question of whether intravenous rtPA given 3 to 6 hours after stroke onset promotes reperfusion and attenuates infarct growth in patients who have a “mismatch” between perfusion-weighted and diffusion-weighted MRI. Intravenous rtPA was nonsignificantly associated with lower infarct growth but significantly associated with increased reperfusion in patients who had mismatch.29,255,286 In the Diffusion-Weighted Imaging Evaluation for Understanding Stroke Evolution (DEFUSE) study, a target mismatch pattern of small core and large penumbra was associated with greater clinical response to reperfusion.345,346,362,363 DEDAS (Dose Escalation of Desmoteplase for Acute Ischemic Stroke)347 appeared to show intravenous desmoteplase to be safe and led to 2 pivotal studies, Desmoteplase in Acute Ischemic Stroke (DIAS) 1 and 2, that tested the concept of using advanced MR or CT for intravenous fibrinolysis triage in the 3- to 9-hour time window.349,364 Unfortunately, there was no clinical benefit demonstrated, although favorable trends were seen in the MR-selected patients.364 Newer studies are under way that incorporate lessons from these experiences.
Conclusions and Recommendations
Brain and vascular imaging remains a required component of the emergency assessment of patients with suspected stroke and TIA. Either CT or MRI may be used as the initial imaging test. MRI is more sensitive to the presence of ischemia, but at most institutions, CT remains the most practical initial brain imaging test. A physician skilled in assessing CT or MRI studies should be available to promptly examine the initial scan. In particular, the scan should be evaluated for evidence of early signs of infarction, vessel thrombosis, or bleed. For ischemic stroke patients, both CT and MRI platforms offer powerful multimodal imaging capabilities. Generally, an institution may adopt one platform as its primary imaging strategy and optimize systems operations to attain rapid and reliable scan performance. For patients with rapidly transient symptoms, diffusion MRI provides unique insight into whether a stroke has occurred and is the preferred modality if available. Information about multimodal CT and MRI of the brain suggests that these diagnostic studies provide important information about the diagnosis, prognosis, and appropriate treatment of patients with acute stroke. Emergency imaging of the intracranial vasculature is particularly useful in those institutions that provide endovascular recanalization therapies.
Recommendations for Patients With Acute Cerebral Ischemic Symptoms That Have Not Yet Resolved
Emergency imaging of the brain is recommended before initiating any specific therapy to treat acute ischemic stroke (Class I; Level of Evidence A). In most instances, NECT will provide the necessary information to make decisions about emergency management. (Unchanged from the previous guideline13)
Either NECT or MRI is recommended before intravenous rtPA administration to exclude ICH (absolute contraindication) and to determine whether CT hypodensity or MRI hyperintensity of ischemia is present (Class I; Level of Evidence A). (Revised from the 2009 imaging scientific statement9)
Intravenous fibrinolytic therapy is recommended in the setting of early ischemic changes (other than frank hypodensity) on CT, regardless of their extent (Class I; Level of Evidence A). (Revised from the 2009 imaging scientific statement9)
A noninvasive intracranial vascular study is strongly recommended during the initial imaging evaluation of the acute stroke patient if either intra-arterial fibrinolysis or mechanical thrombectomy is contemplated for management but should not delay intravenous rtPA if indicated (Class I; Level of Evidence A). (Revised from the 2009 imaging scientific statement9)
In intravenous fibrinolysis candidates, the brain imaging study should be interpreted within 45 minutes of patient arrival in the ED by a physician with expertise in reading CT and MRI studies of the brain parenchyma (Class I; Level of Evidence C). (Revised from the previous guideline13)
CT perfusion and MRI perfusion and diffusion imaging, including measures of infarct core and penumbra, may be considered for the selection of patients for acute reperfusion therapy beyond the time windows for intravenous fibrinolysis. These techniques provide additional information that may improve diagnosis, mechanism, and severity of ischemic stroke and allow more informed clinical decision making (Class IIb; Level of Evidence B). (Revised from the 2009 imaging scientific statement9)
Frank hypodensity on NECT may increase the risk of hemorrhage with fibrinolysis and should be considered in treatment decisions. If frank hypodensity involves more than one third of the MCA territory, intravenous rtPA treatment should be withheld (Class III; Level of Evidence A). (Revised from the 2009 imaging scientific statement9)
Recommendations for Patients With Cerebral Ischemic Symptoms That Have Resolved
Noninvasive imaging of the cervical vessels should be performed routinely as part of the evaluation of patients with suspected TIAs (Class I; Level of Evidence A). (Unchanged from the 2009 TIA scientific statement6)
Noninvasive imaging by means of CTA or MRA of the intracranial vasculature is recommended to exclude the presence of proximal intracranial stenosis and/or occlusion (Class I; Level of Evidence A) and should be obtained when knowledge of intracranial steno-occlusive disease will alter management. Reliable diagnosis of the presence and degree of intracranial stenosis requires the performance of catheter angiography to confirm abnormalities detected with noninvasive testing. (Revised from the 2009 TIA scientific statement6)
Patients with transient ischemic neurological symptoms should undergo neuroimaging evaluation within 24 hours of symptom onset or as soon as possible in patients with delayed presentations. MRI, including DWI, is the preferred brain diagnostic imaging modality. If MRI is not available, head CT should be performed (Class I; Level of Evidence B). (Unchanged from the 2009 TIA scientific statement6)
General Supportive Care and Treatment of Acute Complications
Airway, Ventilatory Support, and Supplemental Oxygen
Stroke is a primary failure of focal tissue oxygenation and energy supply. Thus, it is intuitive that systemic hypoxemia and hypotension be avoided and, if present, corrected to limit further cellular damage. Initial assessment of the airway, breathing, and circulation occurs in the prehospital setting and again on arrival in the ED. Constant reassessment of the airway, breathing, and circulation is required to identify oxygen desaturation, respiratory compromise, and hypotension.
Hypoxia appears frequently after stroke. In one small study of hemiparetic patients, 63% developed hypoxia (defined as oxygen saturation <96% for a period >5 minutes) within 48 hours of stroke onset. In those with a history of cardiac or pulmonary disease, all were noted to develop hypoxemia.365 In another study assessing nocturnal hypoxia in stroke patients, 50% (120 of 238) of potentially eligible subjects were excluded because of oxygen requirements. Of the enrolled patients, one third had a mean nocturnal oxygen saturation <93%, and 6% had a saturation <90%.366
Common causes of hypoxia include partial airway obstruction, hypoventilation, aspiration, atelectasis, and pneumonia. Patients with decreased consciousness or brain stem dysfunction are at increased risk of airway compromise because of impaired oropharyngeal mobility and loss of protective reflexes.367,368 Central periodic breathing (Cheyne-Stokes respirations) is a frequent complication of stroke and is associated with decreases in oxygen saturation.369,370 Given the frequency of hypoxia, careful observation and prevention are essential.
Patient Positioning and Monitoring
Data indicate patient positioning can influence oxygen saturation,371 cerebral perfusion pressure, MCA mean flow velocity,372,373 and intracranial pressure (ICP).373 The ideal position of a stroke patient to optimize these parameters, however, is unknown, and the clinician must balance often competing interests, as well as patient tolerance.
Available evidence suggests that in stroke patients without hypoxia or significant respiratory or pulmonary comorbidities, the supine or side position has minimal effect on oxygen saturation.371,374–377 Limited data suggest stroke patients with hypoxia or significant pulmonary comorbidities have lower oxygen saturation in the supine position than in upright positions.371,377 In patients who are able to maintain oxygenation while lying flat, the supine position may offer advantages in cerebral perfusion.372,373
Thus, in nonhypoxic patients able to tolerate lying flat, a supine position is recommended. Patients at risk for airway obstruction or aspiration and those with suspected elevated ICP378 should have the head of the bed elevated 15° to 30°. When patient position is altered, close monitoring of the airway, oxygenation, and neurological status is recommended, and adjustment to changing clinical parameters may be required.
Although provision of supplemental oxygen may seem intuitive, only limited data exist regarding its benefit. A pilot study found that high-flow, normobaric oxygen, started within 12 hours of stroke onset, may be associated with a transient improvement in neurological impairments379 and improvements in MRI spectroscopy and diffusion/perfusion imaging.380 Another feasibility study, however, found no significant differences in patients with MCA territory infarctions treated with 40% oxygen via Venturi mask compared with oxygen 2 L/min delivered via nasal cannula.381 The results of a large, quasi-randomized controlled trial in stroke found no statistical difference in 1-year mortality or neurological disability between patients who received 3 L of oxygen per minute via nasal cannula for 24 hours after admission and those who received no treatment.382
On the basis of these data, it is not apparent that routine supplemental oxygen is required acutely in nonhypoxic patients with mild or moderate strokes. Supplemental oxygen may be beneficial in patients with severe strokes, although the present data are inconclusive, and further research in this area is recommended.382 On the basis of data from reviews largely focusing on resuscitated post–cardiac arrest patients, recent AHA guidelines for emergency cardiovascular care for stroke and resuscitated cardiac arrest patients recommend administration of oxygen to hypoxemic patients to maintain oxygen saturation >94%.15 When oxygen therapy is indicated, it is reasonable to use the least invasive method possible to achieve normoxia. Available methods include nasal cannula, Venturi mask, nonrebreather mask, bilevel positive airway pressure, continuous positive airway pressure, or endotracheal intubation with mechanical ventilation.
No clinical trial has tested the utility of endotracheal intubation in the management of critically ill patients with stroke. It is generally agreed that endotracheal intubation and mechanical ventilation should be performed if the airway is threatened. Evidence suggests that prevention of early aspiration reduces the incidence of pneumonia,383 and protection of the airway may be an important approach in certain patients. Endotracheal intubation and mechanical ventilation may also assist in the management of elevated ICP or malignant brain edema after stroke.378,384 The need for intubation has prognostic implications. Although a small percentage of patients may have a satisfactory outcome after intubation,385 the overall prognosis of intubated stroke patients is poor, with up to 50% mortality within 30 days after stroke.386–388
Approximately one third of patients admitted with stroke will be hyperthermic (temperature >37.6°C) within the first hours after stroke onset.389,390 In the setting of acute ischemic stroke, hyperthermia is associated with poor neurological outcome, possibly secondary to increased metabolic demands, enhanced release of neurotransmitters, and increased free radical production.389,391–398
The physician should determine the source of hyperthermia. Hyperthermia may be secondary to a cause of stroke, such as infective endocarditis, or may represent a complication, such as pneumonia, urinary tract infection (UTI), or sepsis. Because of the negative effects of hyperthermia, maintenance of normothermia or lowering of an acutely elevated body temperature has been hypothesized to improve the prognosis of patients with stroke.399 Measures to achieve normothermia or prevent hyperthermia include both pharmacological and mechanical interventions.
Sulter et al400 found that either aspirin or acetaminophen was modestly successful in achieving normothermia, but those patients with a temperature >38°C were relatively unresponsive to this treatment. In a small, randomized trial, Kasner et al401 administered 3900 mg of acetaminophen daily to afebrile patients with stroke. They concluded that the medication might prevent hyperthermia or modestly promote hypothermia but that the effects were not likely to have a robust clinical impact. Dippel et al402 tested 2 different doses of acetaminophen in a small clinical trial and concluded a daily dosage of 6000 mg might have a potential beneficial effect in lowering body temperature. In a subsequent study, Dippel et al403 compared the effects of placebo, ibuprofen, or acetaminophen on body temperature and demonstrated that no differences in mean body temperature were observed after 24 hours of treatment.
A large, 2500-patient, randomized, double-blind, placebo-controlled trial evaluating whether early treatment with acetaminophen improved functional outcome by reducing body temperature and fever prevention found no statistical difference between groups; however, the trial was terminated prematurely (after 1400 patients) because of lack of funding.404 Post hoc analysis identified a beneficial effect in patients with a baseline body temperature of 37°C to 39°C; however, this was not a prespecified analysis. Treated patients had a mean body temperature 0.26°C (95% CI, 0.18°C–0.31°C) lower than the control group 24 hours after starting therapy.404 More recently, an updated meta-analysis of the relationship of hyperthermia and stroke mortality in patients with acute stroke demonstrated a 2-fold increase in short-term mortality in patients with hyperthermia within the first 24 hours of hospitalization.398
Although strong experimental and clinical evidence indicates that induced hypothermia can protect the brain in the presence of global hypoxia or ischemia, including after cardiac arrest, data about the utility of induced hypothermia for treatment of patients with stroke are not yet available. Hypothermia is discussed in more detail in the "Neuroprotective Agents" section of this statement.
Cardiac monitoring begins in the prehospital setting and continues throughout the initial assessment and management of acute stroke. As mentioned before, continuous cardiac monitoring is indicated for at least the first 24 hours after stroke.136,405,406 Recent studies have suggested Holter monitoring is more effective in identifying atrial fibrillation or other serious arrhythmias after stroke.134 Outpatient event monitoring may be indicated in patients with cryptogenic stroke and suspected paroxysmal arrhythmias, especially in those patients with short hospitalizations in which monitoring was brief. The utility of prophylactic administration of medications to prevent cardiac arrhythmias among patients with stroke is not known.
Arterial blood pressure is a dynamic parameter that can fluctuate significantly, with clinical consequences. Elevated blood pressure is common during acute ischemic stroke. In one observational study, the systolic blood pressure was >139 mm Hg in 77% and >184 mm Hg in 15% of patients on arrival at the ED.407 The blood pressure is often higher in acute stroke patients with a history of hypertension than in those without premorbid hypertension. Blood pressure typically decreases spontaneously during the acute phase of ischemic stroke, starting within 90 minutes after onset of stroke symptoms.408–414 Extreme arterial hypertension is clearly detrimental, because it leads to encephalopathy, cardiac complications, and renal insufficiency. Theoretically, moderate arterial hypertension during acute ischemic stroke might be advantageous by improving cerebral perfusion of the ischemic tissue, or it might be detrimental by exacerbating edema and hemorrhagic transformation of the ischemic tissue. Extreme arterial hypotension is clearly detrimental, because it decreases perfusion to multiple organs, especially the ischemic brain, exacerbating the ischemic injury. Thus, an arterial blood pressure range likely exists that is optimal during acute ischemic stroke on an individual basis. Unfortunately, such an ideal blood pressure range has not yet been scientifically determined. It is likely that an ideal blood pressure range during acute ischemic stroke will depend on the stroke subtype and other patient-specific comorbidities.
Multiple studies investigated various blood pressure parameters during the admission for acute ischemic stroke and clinical outcomes. Some studies found a U-shaped relation between the admission blood pressure and favorable clinical outcomes, with an optimal systolic blood pressure ranging from 121 to 200 mm Hg and diastolic blood pressure ranging from 81 to 110 mm Hg415–418 among these studies. However, elevated in-hospital blood pressure during acute ischemic stroke has been associated with worse clinical outcomes in a more linear fashion.419–427
Studies analyzing the extent of in-hospital blood pressure fluctuations during acute ischemic stroke found inconsistent associations with clinical outcomes.415,421,422,424,428,429 Three studies found that decreases in blood pressure were associated with poor clinical outcomes.415,421,428 Two studies found no association between blood pressure fluctuations and clinical outcomes.424,429 One study found that decreases in blood pressure were associated with favorable clinical outcome.422 Although these observational studies analyzed data controlling for confounding factors, the blood pressure treatments were not controlled, and it is impossible to ascertain the role of the blood pressure in relation to the outcomes.
One acute ischemic stroke treatment trial, the Intravenous Nimodipine West European Stroke Trial (INWEST),430 set out to test the calcium channel blocker nimodipine as cytoprotective therapy within 24 hours after ischemic stroke onset and found complications related to blood pressure lowering.408 A decrease in blood pressure was associated with intravenous nimodipine therapy and worse clinical outcome at 21 days. Also, a decrease in diastolic blood pressure >10 mm Hg, but not in the systolic pressure, was significantly associated with worse outcome.
A few preliminary randomized trials of blood pressure lowering in acute ischemic stroke have been published.411,413,431 A placebo-controlled randomized trial tested oral nimodipine starting within 48 hours after ischemic stroke onset in 350 patients.413 The systolic and diastolic blood pressures were both significantly lower in the nimodipine group. Functional outcome at 3 months was similar in the 2 treatment groups, but mortality was significantly higher in the nimodipine group. A placebo-controlled randomized trial of therapy with the angiotensin receptor blocker candesartan cilexetil, starting an average of 30 hours after ischemic stroke onset in 342 patients with elevated blood pressure,431 was stopped early. Although blood pressure and the Barthel index score at 3 months were similar in the 2 study groups, patients who received the active drug had significantly lower mortality and fewer vascular events at 12 months. However, a larger efficacy trial (n=2004) of candesartan therapy with a similar study design showed a mean blood pressure reduction of 7/5 mm Hg at day 7 and no improvement in functional outcome.432 Favorable outcomes at 6 months, however, were less likely with candesartan than with placebo (modified Rankin Scale [mRS] score 0–2 in 75% versus 77%; significant by shift analysis [P=0.048]).
A 3-armed randomized trial tested labetalol or lisinopril compared with placebo starting within 36 hours after stroke onset in 179 patients.411 Inclusion of patients with ICH in this trial (14% of the trial patients) obscures the interpretation of results in relation to acute ischemic stroke patients. Over the initial 24 hours, the systolic blood pressure dropped significantly more in the 2 active treatment groups than in the placebo group (21 mm Hg [≈12%] versus 11 mm Hg). Systolic blood pressure over the initial 24 hours compared with placebo dropped significantly more in the lisinopril group (by 14 mm Hg) than in the labetalol group (by 7 mm Hg). The greater blood pressure drops in the active treatment groups were not associated with complications. The primary outcome of death or dependency at 2 weeks was similar in the 2 active treatment groups overall and among patients with ischemic stroke. However, mortality at 3 months was significantly lower in the 2 active treatment groups (9.7%) than with placebo (20.3%, P=0.05).
The Continue or Stop Post-Stroke Antihypertensives Collaborative Study (COSSACS) compared the continuation of antihypertensive therapy to stopping preexisting antihypertensive drugs during acute hospitalization for ischemic stroke.433 Patients were enrolled within 48 hours of stroke onset and the last dose of antihypertensive medication and were maintained in the 2 treatment arms for 2 weeks. The study was terminated prematurely; however, continuation of antihypertensive medications did not reduce 2-week mortality or morbidity and was not associated with 6-month mortality or cardiovascular event rates.
Adding to the complexity and uncertainty of arterial blood pressure management during acute ischemic stroke, small pilot trials have carefully raised the blood pressure in acute ischemic stroke patients without apparent complications. It remains unclear what the risk-benefit ratio is for lowering or raising the blood pressure during acute ischemic stroke. Larger trials with well-defined criteria are needed. At this time, the previous recommendation not to lower the blood pressure during the initial 24 hours of acute ischemic stroke unless the blood pressure is >220/120 mm Hg or there is a concomitant specific medical condition that would benefit from blood pressure lowering remains reasonable.
Some conditions, such as myocardial ischemia, aortic dissection, and heart failure, may accompany acute ischemic stroke and may be exacerbated by arterial hypertension. When blood pressure management is indicated for a specific medical condition in the setting of concurrent acute cerebral ischemia, an optimal approach has not been determined, and at present, blood pressure targets are based on best clinical judgment. A reasonable estimate might be to initially lower the systolic blood pressure by 15% and monitor for neurological deterioration related to the pressure lowering.
Specific blood pressure management recommendations have been established for acute ischemic stroke patients being considered for fibrinolytic therapy (Table 9). These recommendations include a gentle approach to bringing the pressure below 185/110 mm Hg to qualify for fibrinolytic therapy with intravenous rtPA. Once intravenous rtPA is given, the blood pressure must be maintained below 180/105 mm Hg to limit the risk of ICH. A recently published observational study of 11 080 patients with acute ischemic stroke treated with intravenous rtPA further supports the association between elevated blood pressure and adverse outcomes in this setting.434 Higher blood pressures during the initial 24 hours were associated with greater risk of sICH in a linear fashion. However, a U-shaped relation was found between blood pressure during the initial 24 hours and death or dependency at 3 months, with best outcomes associated with systolic blood pressures of 141 to 150 mm Hg.
Because arterial blood pressure is a dynamic parameter, it is important to monitor it frequently, especially during the first day of stroke, to identify trends and extreme fluctuations that would require intervention. When lowering the blood pressure during acute ischemic stroke is indicated, risk would be minimized by lowering the pressure in a well-controlled manner. Controlled blood pressure lowering during acute stroke can best be achieved with intravenous antihypertensive therapies. A single optimal medication to lower the blood pressure in all patients with acute stroke has not been determined, and an individualized approach is best.
It is reasonable to temporarily discontinue or reduce (to prevent the rare occurrence of antihypertensive withdrawal syndrome, primarily seen in β-blocker discontinuation) premorbid antihypertensive medications at the onset of acute ischemic stroke, because swallowing is often impaired, and responses to the medications may be less predictable during the acute stress.435 The optimal time after the onset of acute ischemic stroke to restart or start long-term antihypertensive therapy has not been established. The optimal time may depend on various patient and stroke characteristics. Nonetheless, it is reasonable to initiate long-term antihypertensive therapy after the initial 24 hours from stroke onset in most patients.411 An optimal long-term antihypertensive therapy for patients after stroke has not been definitively established, and it might be best to individualize such therapy based on relevant comorbidities, ability to swallow, and likelihood to continue with the prescribed therapy.
Arterial hypotension is rare during acute ischemic stroke and suggests another cause, such as cardiac arrhythmia or ischemia, aortic dissection, or shock. In a study of 930 patients with acute ischemic stroke, the admission systolic blood pressure was <100 mm Hg in only 2.5% of the patients, and this was associated with ischemic heart disease.412 In a study of 11 080 patients treated with intravenous rtPA for acute ischemic stroke, the admission systolic blood pressure was <100 mm Hg in only 64 (0.6%) of the patients.434 The brain is especially vulnerable to arterial hypotension during acute ischemic stroke because of impaired cerebral autoregulation. Arterial hypotension on admission in acute ischemic stroke patients has been associated with poor outcomes in multiple studies.412,415–417,434 The exact definition of arterial hypotension needs to be individualized. In a given patient, a blood pressure that is lower during acute ischemic stroke than the premorbid pressure could be considered hypotension. Urgent evaluation, diagnosis, and correction of the cause of arterial hypotension are needed to minimize the extent of brain damage. If the arterial hypotension cannot be corrected rapidly by other means, use of vasopressor agents is reasonable. Relatively small trials have evaluated the use of drug-induced hypertension and intravascular volume expansion in acute ischemic stroke, and these are summarized in the “Volume Expansion, Vasodilators, and Induced Hypertension” section of this guideline.
Patients presenting with acute ischemic stroke are predominantly either euvolemic or hypovolemic. Hypovolemia may predispose to hypoperfusion and exacerbate the ischemic brain injury, cause renal impairment, and potentiate thrombosis. Hypervolemia may exacerbate ischemic brain edema and increase stress on the myocardium. Thus, euvolemia is desirable. One observational study found an association between elevated osmolality (>296 mOsm/kg) during the initial 7 days of acute stroke (90% ischemic) and mortality within 3 months after adjustment for potential confounding factors.436 In that study, serum sodium and urea measurements were associated with the measured plasma osmolality and thus might be useful in monitoring hydration status. However, the cause-and-effect relationship between hydration during acute ischemic stroke and outcome remains unclear.
For patients who are euvolemic at presentation, clinicians should initiate maintenance intravenous fluids. Apart from unusual losses, daily fluid maintenance for adults can be estimated as 30 mL per kilogram of body weight.437 For patients who are hypovolemic at presentation, rapid replacement of the depleted intravascular volume followed by maintenance intravenous fluids is reasonable. Although plasma osmolality was similar in acute stroke patients hydrated orally or intravenously,436 some stroke patients have impaired swallowing. Extra precaution is needed in patients who are especially vulnerable to intravascular volume overload, such as those with renal or heart failure. Treatment of patients with specific conditions, such as syndrome of inappropriate antidiuretic hormone secretion or fever, requires modifications to standard hydration protocols.
A substantial proportion of hypotonic solutions, such as 5% dextrose (after the glucose is metabolized) or 0.45% saline, is distributed into the intracellular spaces and may exacerbate ischemic brain edema. Isotonic solutions such as 0.9% saline are more evenly distributed into the extracellular spaces (interstitial and intravascular) and may be better for patients with acute ischemic stroke.
Hypoglycemia during acute ischemic stroke is rare and likely related to antidiabetic medications. If severe enough, hypoglycemia is known to cause autonomic and neurological symptoms, including stroke mimics and seizures. Such symptoms are readily reversible if the hypoglycemia is rapidly corrected. However, if untreated, severe or prolonged hypoglycemia can result in permanent brain damage. Thus, blood glucose should be measured as soon as possible in patients with acute ischemic stroke; low levels (<60 mg/dL) should be corrected urgently.
The combination of symptoms attributable to hypoglycemia and the threshold for such symptoms vary considerably between individuals. In healthy people, autonomic symptoms (such as sweating, trembling, or anxiety) usually begin to appear when the blood glucose level drops below 57 mg/dL, and manifestations of brain dysfunction (such as disorientation, dizziness, or slowing of speech) usually begin to appear when the glucose level drops below 47 mg/dL.438,439 However, in patients with poorly controlled diabetes mellitus, these thresholds are shifted to higher blood glucose levels.438 Occasionally, brain dysfunction occurs before the autonomic symptoms. Hypoglycemia (blood glucose level <60 mg/dL) can be corrected rapidly in most patients with a slow intravenous push of 25 mL of 50% dextrose. Oral glucose–containing solutions are also reasonable treatment options but take longer to raise the blood glucose level and may not be feasible in patients with dysphagia.
Hyperglycemia is common during acute ischemic stroke. Several studies have shown admission blood glucose is elevated in >40% of patients with acute ischemic stroke, most commonly among patients with a history of diabetes mellitus.440,441 Blood glucose elevations during acute stroke are related in part to a nonfasting state and in part to a stress reaction with impaired glucose metabolism. Multiple observational studies have found an association between admission and in-hospital hyperglycemia and worse clinical outcomes than with normoglycemia.442,443 Among stroke patients treated with intravenous rtPA, hyperglycemia has been associated with sICH and worse clinical outcomes.444–447 Also, multiple studies found an association between acute ischemic stroke hyperglycemia and worse outcomes defined by MRI infarct volume.448–451 Although multiple observational studies consistently found an association between acute stroke hyperglycemia and worse outcomes, it cannot be determined whether this is a cause-and-effect relationship on the basis of such studies.
So far, only 1 randomized efficacy trial of hyperglycemia treatment in acute stroke has been reported (the Glucose-Insulin-Stroke Trial–UK [GIST-UK]).452 Patients (n=933) with acute ischemic stroke within 24 hours of symptom onset, not previously treated with insulin, were randomized to unblinded intravenous treatment with insulin, potassium, and glucose versus saline. Protocol treatment continued for 24 hours. Although the results of this trial were neutral (no difference in clinical outcomes between the 2 treatment groups), the design was such that key questions remain unanswered. First, the GIST-UK trial was stopped early, because 2355 subjects were originally planned, and it was thus underpowered to detect a possible treatment effect. Second, the mean glucose level in the insulin-treated group was only 10 mg/dL lower than in the saline control group, and the control group was only mildly hyperglycemic (≈122 mg/dL between hours 8–24). This was likely because of the inclusion of predominantly nondiabetic patients (84%). Larger decreases in glucose levels may be needed to detect a therapeutic effect. Third, the median time to initiation of protocol treatment was 13 hours. Although the optimal time to correct hyperglycemia during acute ischemic stroke has not been established, earlier treatment may have been therapeutic. Pilot clinical trials have demonstrated the feasibility and safety of rapid reductions in glucose levels with intravenous insulin during acute ischemic stroke.453–456 Thus, the definitive efficacy and safety of earlier and greater reductions in glucose levels during acute ischemic stroke remain to be studied.
There is currently no clinical evidence that targeting the blood glucose to a particular level during acute ischemic stroke will improve outcomes. The main risk from aggressive hyperglycemia correction in acute stroke appears to be possible hypoglycemia. Avoidance of hypoglycemia requires frequent glucose monitoring, and in many hospitals this necessitates admission to an intensive care unit, which may otherwise not be needed.
Further clinical trials should establish the efficacy and the risk-benefit ratio of rapid hyperglycemia correction during acute stroke. Also, if lowering hyperglycemia during acute ischemic stroke proves beneficial, it would be useful to know whether this is a linear effect and what glucose levels can be considered dangerously low. In the meantime, it is prudent to treat hyperglycemia during acute stroke in a manner that avoids excessive resources, labor, and risk. It is reasonable to follow the current American Diabetes Association recommendation to maintain the blood glucose in a range of 140 to 180 mg/dL in all hospitalized patients.457 There are multiple subcutaneous and intravenous insulin protocols that use insulin to lower hyperglycemia during hospitalization, and these have not been compared with each other in acute stroke patients. The subcutaneous insulin protocols can safely lower and maintain blood glucose levels below 180 mg/dL in acute stroke patients without excessive use of healthcare resources.453,454,458 However, some hospitals may be prepared to safely administer intravenous insulin to patients with acute stroke and hyperglycemia and maintain the glucose levels considerably below 200 mg/dL.
Cardiac monitoring is recommended to screen for atrial fibrillation and other potentially serious cardiac arrhythmias that would necessitate emergency cardiac interventions. Cardiac monitoring should be performed for at least the first 24 hours (Class I; Level of Evidence B). (Revised from the previous guideline13)
Patients who have elevated blood pressure and are otherwise eligible for treatment with intravenous rtPA should have their blood pressure carefully lowered (Table 9) so that their systolic blood pressure is <185 mm Hg and their diastolic blood pressure is <110 mm Hg (Class I; Level of Evidence B) before fibrinolytic therapy is initiated. If medications are given to lower blood pressure, the clinician should be sure that the blood pressure is stabilized at the lower level before beginning treatment with intravenous rtPA and maintained below 180/105 mm Hg for at least the first 24 hours after intravenous rtPA treatment. (Unchanged from the previous guideline13)
Airway support and ventilatory assistance are recommended for the treatment of patients with acute stroke who have decreased consciousness or who have bulbar dysfunction that causes compromise of the airway (Class I; Level of Evidence C). (Unchanged from the previous guideline13)
Supplemental oxygen should be provided to maintain oxygen saturation >94% (Class I; Level of Evidence C). (Revised from the previous guideline13)
Sources of hyperthermia (temperature >38°C) should be identified and treated, and antipyretic medications should be administered to lower temperature in hyperthermic patients with stroke (Class I; Level of Evidence C). (Unchanged from the previous guideline13)
Until other data become available, consensus exists that the previously described blood pressure recommendations should be followed in patients undergoing other acute interventions to recanalize occluded vessels, including intra-arterial fibrinolysis (Class I; Level of Evidence C). (Unchanged from the previous guideline13)
In patients with markedly elevated blood pressure who do not receive fibrinolysis, a reasonable goal is to lower blood pressure by 15% during the first 24 hours after onset of stroke. The level of blood pressure that would mandate such treatment is not known, but consensus exists that medications should be withheld unless the systolic blood pressure is >220 mm Hg or the diastolic blood pressure is >120 mm Hg (Class I; Level of Evidence C). (Revised from the previous guideline13)
Hypovolemia should be corrected with intravenous normal saline, and cardiac arrhythmias that might be reducing cardiac output should be corrected (Class I; Level of Evidence C). (Revised from the previous guideline13)
Hypoglycemia (blood glucose <60 mg/dL) should be treated in patients with acute ischemic stroke (Class I; Level of Evidence C). The goal is to achieve normoglycemia. (Revised from the previous guideline13)
Evidence from one clinical trial indicates that initiation of antihypertensive therapy within 24 hours of stroke is relatively safe. Restarting antihypertensive medications is reasonable after the first 24 hours for patients who have preexisting hypertension and are neurologically stable unless a specific contraindication to restarting treatment is known (Class IIa; Level of Evidence B). (Revised from the previous guideline13)
No data are available to guide selection of medications for the lowering of blood pressure in the setting of acute ischemic stroke. The antihypertensive medications and doses included in Table 9 are reasonable choices based on general consensus (Class IIa; Level of Evidence C). (Revised from the previous guideline13)
Evidence indicates that persistent in-hospital hyperglycemia during the first 24 hours after stroke is associated with worse outcomes than normoglycemia, and thus, it is reasonable to treat hyperglycemia to achieve blood glucose levels in a range of 140 to 180 mg/dL and to closely monitor to prevent hypoglycemia in patients with acute ischemic stroke (Class IIa; Level of Evidence C). (Revised from the previous guideline13)
The management of arterial hypertension in patients not undergoing reperfusion strategies remains challenging. Data to guide recommendations for treatment are inconclusive or conflicting. Many patients have spontaneous declines in blood pressure during the first 24 hours after onset of stroke. Until more definitive data are available, the benefit of treating arterial hypertension in the setting of acute ischemic stroke is not well established (Class IIb; Level of Evidence C). Patients who have malignant hypertension or other medical indications for aggressive treatment of blood pressure should be treated accordingly. (Revised from the previous guideline13)
Supplemental oxygen is not recommended in nonhypoxic patients with acute ischemic stroke (Class III; Level of Evidence B). (Unchanged from the previous guideline13)
Intravenous fibrinolytic therapy for acute stroke is now widely accepted.459–467 The US FDA approved the use of intravenous rtPA in 1996, in part on the basis of the results of the 2-part NINDS rtPA Stroke Trial, in which 624 patients with ischemic stroke were treated with placebo or intravenous rtPA (0.9 mg/kg IV, maximum 90 mg) within 3 hours of symptom onset, with approximately one half treated within 90 minutes.166 In the first trial (Part I), the primary end point was neurological improvement at 24 hours, as indicated by complete neurological recovery or an improvement of 4 points on the NIHSS. In the second trial (Part II), the pivotal efficacy trial, the primary end point was a global OR for a favorable outcome, defined as complete or nearly complete neurological recovery 3 months after stroke. Treatment with intravenous rtPA was associated with an increase in the odds of a favorable outcome (OR, 1.9; 95% CI, 1.2–2.9). Excellent outcomes on individual functional measures were more frequent with intravenous rtPA for global disability (40% versus 28%), global outcome (43% versus 32%), activities of daily living (53% versus 38%), and neurological deficits (34% versus 20%). The benefit was similar 1 year after stroke.468
The major risk of intravenous rtPA treatment remains sICH. In the NINDS rtPA Stroke Trial, early minimal neurological symptoms or neurological deterioration temporally associated with any intracranial hemorrhage occurred in 6.4% of patients treated with intravenous rtPA and 0.6% of patients given placebo. However, mortality in the 2 treatment groups was similar at 3 months (17% versus 20%) and 1 year (24% versus 28%).166,469 Although the presence of edema or mass effect on baseline CT scan was associated with higher risk of sICH, patients with these findings were more likely to have an excellent outcome if they received fibrinolytic therapy.470 The presence of early ischemic changes on CT scan was not associated with adverse outcome.148 The likelihood of a favorable outcome also was associated with the severity of deficits and the patient’s age. Patients with mild to moderate strokes (NIHSS score <20) and people <75 years of age had the greatest potential for an excellent outcome with treatment.103 The chances of a complete or nearly complete recovery among patients with severe stroke (NIHSS score of >20) improved with treatment, but such recovery occurred less often in this group of critically ill patients.103 Four subsequent trials, the European Cooperative Acute Stroke Study (ECASS I and ECASS II) and the Alteplase Thrombolysis for Acute Noninterventional Therapy in Ischemic Stroke (ATLANTIS A and ATLANTIS B), enrolled subsets of patients in the ≤3-hour time period and found largely similar effects in this time window to those observed in the 2 NINDS rtPA trials.92,167,462,471–473
Debate about time of initiation of intravenous rtPA treatment merits attention. The NINDS investigators reported a time-to-treatment interaction in a subgroup analysis of the NINDS rtPA Stroke Trial.93 Treatment with intravenous rtPA initiated within 90 minutes of symptom onset was associated with an OR of 2.11 (95% CI, 1.33–3.55) for favorable outcome at 3 months compared with placebo. In comparison, the OR for good outcome at 3 months for treatment with intravenous rtPA initiated within 90 to 180 minutes was 1.69 (95% CI, 1.09–2.62). The investigators concluded that the earlier that treatment is initiated, the better the result. A subsequent pooled analysis of all large, multicenter, placebo-controlled trials of intravenous rtPA for acute stroke confirmed a time effect.468 Investigation of the early time epoch in the NINDS trials revealed a potential confounder in the original data: 19% of the patients treated with intravenous rtPA between 91 and 180 minutes after stroke onset had an NIHSS score of <5 compared with 4% of the placebo patients. On the basis of this observation, it has been suggested that the relative preponderance of mild strokes with a likely good outcome in the intravenous rtPA treatment group may explain the entire benefit reported for patients treated between 91 and 180 minutes. Subsequent reanalysis showed that the imbalance in patients with minor stroke did not explain the difference between treatment and placebo.474 The adjusted OR for 3-month favorable outcome (ORs for treatment compared with placebo) for the subgroup of patients from the 2 NINDS intravenous rtPA stroke trials with NIHSS score of <5 at baseline and time from stroke onset to treatment of 91 to 180 minutes was statistically significant in favor of treatment. Indeed, when all possible subgroups were examined separately, no effect of the severity imbalance could be shown to influence the overall result that intravenous rtPA therapy positively influenced outcome. In separate analyses by independent groups, an identical finding was reached: Baseline imbalances in the numbers of patients with mild stroke did not explain the overall study result.475–477
Subsequent to the approval of intravenous rtPA for treatment of patients with acute ischemic stroke, numerous groups reported on the utility of the treatment in a community setting.117,120,122,478–483 Some groups reported rates of intracranial hemorrhage and favorable outcomes that were similar to those found in the NINDS trials, but others did not. It is now clear that the risk of hemorrhage is proportional to the degree to which the NINDS protocol is not followed.120,483,484 In addition to the risk of sICH, other potential adverse experiences include systemic bleeding, myocardial rupture if fibrinolytics are given within a few days of acute myocardial infarction, and reactions such as anaphylaxis or angioedema, although these events are rare.460
Orolingual angioedema reactions (swelling of tongue, lips, or oropharynx) are typically mild, transient, and contralateral to the ischemic hemisphere.485 Angioedema is estimated to occur in 1.3% to 5.1% of all patients who receive intravenous rtPA treatment for ischemic stroke.464,485,486 Risk of angioedema is associated with angiotensin-converting enzyme inhibitor use and with infarctions that involve the insular and frontal cortex. Empiric monitoring recommendations include inspection of tongue, lips, and oropharynx after intravenous rtPA administration. Empiric treatment recommendations include intravenous ranitidine, diphenhydramine, and methylprednisolone.486
The largest community experience, the SITS-ISTR Registry (Safe Implementation of Thrombolysis in Stroke–International Stroke Thrombolysis Register, which incorporates the SITS-MOST [Safe Implementation of Thrombolysis in Stroke–Monitoring Study] Registry), resulted when, in 2002, the European Medicines Evaluation Agency granted license for the use of intravenous rtPA for the treatment of ischemic stroke patients within 3 hours of symptom onset. The approval was conditional on the completion of a prospective registry of patient treatment experience with intravenous rtPA within the 3-hour window from stroke onset. SITS-ISTR reported on 11 865 patients treated within 3 hours of onset at 478 centers in 31 countries worldwide.468 The frequency of early neurological deterioration temporally associated with substantial parenchymal hematoma after intravenous rtPA was 1.6% (95% CI, 1.4%–1.8%). The frequency of favorable outcome (combined mRS scores of 0, 1, and 2) at 90 days was 56.3% (CI, 55.3%–57.2%) in the intravenous rtPA patients, comparable to the favorable outcome rate among patients treated within 3 hours in the pooled analysis of the 6 randomized trials.468 These findings appear to confirm the safety of intravenous rtPA within the 3-hour window at sites that have an institutional commitment to acute stroke care.
With >15 years of fibrinolytic experience in acute ischemic stroke, multiple groups have reported their outcomes in treating patients with “off-label” fibrinolysis.487–493 These groups report the use of fibrinolysis in patients with conditions including extreme age (>80 years), prior stroke and diabetes mellitus, minor stroke, rapidly improving stroke symptoms, recent myocardial infarction, major surgery or trauma within the preceding 3 months, and oral anticoagulation use. Overall, the outcomes in the treated patients with these contraindications were better than nontreated “controls” from registry data. Rates of sICH were not increased in these reports. Because stroke patients continue to present with conditions not specifically stated in the original indications for and usage of intravenous rtPA, further experience may allow consideration for fibrinolysis in these situations.
Extended Window for Intravenous rtPA
Subsequent to the NINDS trials, 5 clinical trials have tested the use of intravenous rtPA up to 6 hours after stroke onset without specialized imaging for patient selection. The first 4 trials, ECASS I, ECASS II, ATLANTIS A, and ATLANTIS B,167,471,473,494 collectively enrolled 1847 patients in the 3- to 6-hour time period. None of these 4 trials was individually positive on its prespecified primary end point. In a pooled individual patient-level analysis of these 4 trials, a benefit of therapy in the 3- to 4.5-hour window was suggested, both in increasing the rate of excellent outcomes (adjusted OR, 1.40; 95% CI, 1.05–1.85) and in improving outcomes along the entire range of poststroke disability.92,495 Fibrinolytic therapy in the 4.5- to 6-hour window produced a statistically nonsignificant increase in the rate of excellent outcomes (adjusted OR, 1.15; 95% CI, 0.90–1.47).92,495 In the 3- to 4.5-hour window, across all trials, rates of radiological parenchymal hematoma were higher with fibrinolytic therapy, 5.9% versus 1.7%, but mortality was not increased at 13% versus 12%. In the 4.5- to 6-hour window, fibrinolytic therapy increased rates of both radiological parenchymal hematoma (6.9% versus 1.0%) and mortality (15% versus 10%). When data from all time windows in the first 6 large intravenous rtPA trials were pooled, a time-to-treatment interaction was shown.92 Treatment with intravenous rtPA initiated within 1.5 hours of symptom onset was associated with an OR of 2.81 (95% CI, 1.75–4.50) for favorable outcome at 3 months compared with placebo. The OR for good outcome at 3 months for treatment with intravenous rtPA initiated within 1.5 to 3 hours was 1.55 (95% CI, 1.12–2.15) compared with 1.40 (95% CI, 1.05–1.85) within 3 to 4.5 hours and 1.15 (0.90–1.47) within 4.5 to 6 hours.
The ECASS III trial was undertaken to prove or disprove the benefit of intravenous rtPA in the 3- to 4.5-hour window suggested by the pooled analysis of the 4 prior trials. In ECASS III, patients between 3.0 and 4.5 hours from symptom onset were randomized to either intravenous rtPA (n=418) or placebo (n=403).169 The dosing regimen was 0.9 mg/kg (maximum of 90 mg), with 10% given as an initial bolus and the remainder infused over 1 hour.13 The inclusion and exclusion criteria for the trial were similar to those in the existing AHA Stroke Council guidelines for treatment of patients within 3 hours of stroke onset,13 except for the time window and that the trial additionally excluded people >80 years old, those with a baseline NIHSS score >25, those taking oral anticoagulants (even if their INR was <1.7), and those who had the combination of a previous stroke and diabetes mellitus. Patients were permitted to receive low-dose parenteral anticoagulants for prophylaxis of DVT within 24 hours after treatment with intravenous rtPA.
Early neurological deterioration likely caused by intracranial hemorrhage was identified in 10 subjects treated with intravenous rtPA (2.4%) and 1 subject administered placebo (0.2%; OR, 9.85; 95% CI, 1.26–77.32; P=0.008).169 However, mortality in the 2 treatment groups did not differ significantly and was nominally higher among the subjects treated with placebo.169 The primary efficacy outcome in ECASS III was excellent 90-day outcome on the mRS global disability scale (mRS score 0–1). This outcome was more frequent with intravenous rtPA (52.4%) than placebo (45.2%; OR, 1.34; 95% CI, 1.02–1.76; risk ratio, 1.16; 95% CI, 1.01–1.34; P=0.04). The ECASS III findings align with preclinical and clinical data that suggest a time dependency for benefit from treatment with intravenous rtPA. The point estimate for the degree of benefit seen in ECASS III (OR for global favorable outcome, 1.28; 95% CI, 1.00–1.65) was less than the point estimate of that found in the pool of patients enrolled from 0 to 3 hours in the NINDS study (OR, 1.9; 95% CI, 1.2–2.9)166,169 and was similar to the pooled analysis of the results of subjects enrolled in the 3- to 4.5-hour window in previous trials of intravenous rtPA (OR, 1.4).92,166,167,471,473,494 Overall, the ECASS III results were consistent with the results of previous trials,92,496,497 which indicates that intravenous rtPA can be given safely to, and can improve outcomes for, carefully selected patients treated 3 to 4.5 hours after stroke.
In June 2012, the results from the Third International Stroke Trial (IST-3), the largest randomized, placebo-controlled trial to date of intravenous rtPA, were published.498 The trial enrolled 3035 patients who were randomized to treatment within 6 hours from symptom onset with 0.9 mL/kg in the active arm. Eligibility criteria were similar to other intravenous rtPA trials with several exceptions, including no upper limit to age and broader blood pressure eligibility (systolic blood pressure 90–220 mm Hg and diastolic blood pressure 40–130 mm Hg). The primary outcome measure, an Oxford Handicap Score of 0 to 2 (alive and independent) at 6 months, was achieved in 37% of patients in the intravenous rtPA group versus 35% in the control group (OR, 1.13; 95% CI, 0.95–1.35; P=0.181). Using an ordinal analysis, there was a significant shift in overall Oxford Handicap Score (OR, 1.27; 95% CI, 1.10–1.47; P=0.001). Within 7 days, fatal or nonfatal sICH occurred in 7% versus 1% in the treatment versus placebo arms, respectively. More deaths occurred within 7 days in the intravenous rtPA group (11%) than in the control group (7%; adjusted OR, 1.60; 95% CI, 1.22–2.08; P=0.001), but by 6 months, 27% of patients had died in both groups.
Also in June 2012, Sandercock and colleagues498 published a meta-analysis of 12 intravenous rtPA trials that had enrolled 7012 patients up to 6 hours from symptom onset. The results confirmed the benefits of intravenous rtPA administered within 6 hours from symptom onset, with final follow-up mRS score of 0 to 2 in 46.3% of intravenous rtPA–treated patients compared with 42.1% of patients in the placebo arms (OR, 1.17; 95% CI, 1.06–1.29; P=0.001). The data also reinforced the importance of timely treatment, because the benefit of intravenous rtPA was greatest in patients treated within 3 hours from symptom onset (mRS score 0–2, 40.7% versus 31.7%; OR, 1.53, 95% CI, 1.26–1.86; P<0.0001). As noted in the IST-3 trial, sICH events were more common in the intravenous rtPA group (7.7% versus 1.8%; OR, 3.72, 95% CI, 2.98–4.64; P<0.0001), and death within 7 days was increased in intravenous rtPA patients (8.9%) compared with the placebo arms (6.4%; OR, 1.44, 95% CI, 1.18–1.76; P=0.0003), but by final follow-up, the number of deaths was similar (19.1% versus 18.5%; OR, 1.06, 95% CI, 0.94–1.20; P=0.33). Importantly, the authors found patients of all ages received benefit from intravenous rtPA treatment compared with placebo.
Drug regulatory authorities have recently taken contradictory actions with regard to later administration of intravenous rtPA, with the European Medicines Agency expanding approval of intravenous rtPA to the 3- to 4.5-hour window and the US FDA declining to do so. The basis of these decisions currently remains confidential as part of the regulatory process. To inform this update of the guidelines, the AHA/ASA Writing Committee leadership requested and was granted by the US manufacturer (Genentech) partial access to the FDA decision correspondence. The degree of evidence that AHA/ASA requires for a Grade B recommendation is less than for a Grade A recommendation, and the latter generally more closely approximates the level of evidence that the FDA requires for label approval. On the basis of the review, it is the opinion of the writing committee leadership that the existing Grade B recommendation remains reasonable. The sponsor indicated it planned to work with academic investigators to independently replicate the types of analyses undertaken as part of the FDA review process and make the resultant findings public, and this approach was supported by the writing committee.
Although the maximum time window in which fibrinolytic therapy may be given in many patients has been expanded to 4.5 hours, preclinical, cerebrovascular imaging, and clinical trial evidence indicate the fundamental importance of minimizing total ischemic time and restoring blood flow to threatened but not yet infarcted tissue as soon as feasible. Experience with acute myocardial infarction and acute ischemic stroke systems of care have demonstrated that health system responsiveness is improved by the establishment and monitoring of a time interval within which most patients should be treated after first presentation to the hospital.499,500 Health systems should set a goal of increasing their percentage of stroke patients treated within 60 minutes of presentation to hospital (door-to-needle time of 60 minutes) to at least 80%.43,501,502
Patients With Minor and Isolated or Rapidly Improving Neurological Signs
Minor and isolated symptoms are those that are not presently potentially disabling. Although most patients with potentially disabling symptoms will have NIHSS scores ≥4, certain patients, such as those with gait disturbance, isolated aphasia, or isolated hemianopia, may have potentially disabling symptoms although their NIHSS score is just 2.
Several studies have now reported that approximately one third of patients who are not treated with intravenous rtPA because of mild or rapidly improving stroke symptoms on hospital arrival have a poor final stroke outcome.503–507 A persistent large-artery occlusion on imaging, despite minor symptoms or clinical improvement, may identify patients at increased risk of subsequent deterioration.508 In light of these observations, the practice of withholding intravenous fibrinolytic therapy because of mild or rapidly improving symptoms has been questioned, which justifies further study.
Patients Taking Direct Thrombin Inhibitors and Direct Factor Xa Inhibitors
New classes of anticoagulants are rapidly changing the way physicians treat and prevent disorders of thrombosis. Although most potential agents are in clinical development, the direct thrombin inhibitor dabigatran and the direct factor Xa inhibitor rivaroxaban have been approved for use in the United States. Other factor Xa inhibitors are on the horizon: Apixiban has recently been approved by the FDA, and edoxaban is in the late stages of clinical development. These classes of oral anticoagulants do not require therapeutic monitoring, have fewer side effects (especially lower rates of major hemorrhage), and have fewer drug and food interactions than warfarin.509–512 The challenge for physicians evaluating and considering treatment options for patients with acute ischemic stroke is determining the anticoagulant effect of these agents and estimating the potential increased risk of hemorrhage after reperfusion strategies are initiated.
Specific to dabigatran, drug concentrations peak ≈2 to 3 hours after an oral dose. The active moiety has a half-life of 12 to 17 hours and is cleared primarily by renal elimination. In patients with impaired renal function, the half-life may extend to 20 to 30 hours. The challenge for the physician treating acute stroke patients with this agent is estimating the impact of the drug on the coagulation system. Traditional coagulation tests are not reliable for measuring the anticoagulant effect of dabigatran. The effects of dabigatran on the INR are not predictable. Similarly, the effects of dabigatran on aPTT are not predictable. Although there is correlation between dabigatran plasma concentrations and aPTT results, the correlation is nonlinear. TT and ECT both show a good linear correlation with direct thrombin inhibitors, including dabigatran, and are very sensitive. If the TT or ECT is normal, it is reasonable to assume that plasma concentrations of dabigatran are minimal. Regrettably, these tests are not performed routinely in the ED, and results may take hours to become available.
Specific to the direct factor Xa inhibitors, rivaroxaban has a half-life of 5 to 9 hours and is cleared by renal, fecal, and hepatic mechanisms, whereas apixaban has a half-life of 8 to 15 hours and is cleared by the cytochrome P450 system. The direct factor Xa inhibitors may cause prolongation of the PT and aPTT, but these indexes are not reliable for measuring the pharmacodynamics effects of these agents. Direct factor Xa activity assays may be able to indicate treatment effects but are not routinely performed in the ED, and results may take hours to become available.
Until a simple, fast, and reliable method is determined to measure the clinical impact of the direct thrombin inhibitors and direct factor Xa inhibitors, and more data are collected on use of fibrinolytics and reperfusion strategies in patients taking these classes of drugs, a good medical history will be critical. In patients known to have taken one of these agents in the past, but for whom history or a readily available assay suggests no current substantial anticoagulant effects of the agent, cautious treatment may be pursued. In patients with historical or assay suggestion of at least modest anticoagulant effects of dabigatran, fibrinolytic therapy is likely to be of greater risk and ordinarily would not be undertaken. As other classes of anticoagulants become available for clinical use, similar considerations will be necessary.
For instance, as this guideline was undergoing revisions, the results of 2 large phase III trials of oral direct factor Xa inhibitors for the treatment of patients with atrial fibrillation were published.513,514 These medications, rivaroxaban (FDA approved) and apixaban (recently approved), are pharmacologically different from dabigatran. The recommendations made for dabigatran may not be applicable in all cases for these newer agents because of differences in metabolism. We urge caution in applying these recommendations to these new oral direct factor Xa inhibitor agents.
Other Fibrinolytic Agents
Clinical trials of streptokinase (administered at the treatment dose for acute myocardial ischemia, 1.5 million units) were halted prematurely because of unacceptably high rates of hemorrhage, and this agent should not be used.515–518 Other intravenously administered fibrinolytic agents, including reteplase, urokinase, anistreplase, and staphylokinase, have not been tested extensively. Tenecteplase is a modified tissue plasminogen activator with a longer half-life and higher fibrin specificity than alteplase and appears promising as an effective fibrinolytic, with greater reperfusion and major vessel recanalization with fewer bleeding complications than alteplase in pilot studies. Recently, a US phase IIb study of intravenous tenecteplase in acute stroke was terminated prematurely for nonsafety issues, but an Australian phase IIb trial comparing tenecteplase with alteplase showed significantly improved rates of reperfusion and clinical outcomes by use of imaging-based patient selection.519–521
Desmoteplase is a fibrinolytic agent isolated from vampire bat saliva. Two phase II trials of desmoteplase provided encouraging safety and potential efficacy data in penumbral imaging–selected patients 9 hours after stroke onset.347,349 However, a larger trial revealed no benefit of either of 2 doses of desmoteplase over placebo, possibly because of a higher than projected good outcome rate in the placebo group. Phase III studies are ongoing.
Extracts derived from pit viper venom have been demonstrated to cleave fibrinogen rather than fibrin, reducing plasma fibrinogen, which leads to reduced blood viscosity, increased blood flow, and the prevention of clot formation and/or clot extension. Ancrod, a defibrinogenating agent, has been investigated in patients with acute ischemic stroke.522–526 A systematic meta-analysis of defibrinogenating agents in acute ischemic stroke analyzed 6 trials involving 4148 subjects. The review authors identified a trend toward benefit in reducing death or dependency at the end of the follow-up period (43.7% versus 46.7%, for an absolute risk reduction of 3.0% [95% CI, −0.1% to 5.9%]). The meta-analysis also found that treatment increased early minimal neurological symptoms or neurological deterioration temporally associated with any intracranial hemorrhage (4.9% versus 1.0%, for an absolute risk increase of 3.8% [95% CI, 2.3% to 5.4%]). However, more recently, 2 phase III ancrod trials investigating a refined dosing regimen were stopped after a planned interim analysis found no clinically meaningful difference in outcome between the 2 treatment groups in averting disability.527
Transcranial Ultrasound Fibrinolysis Augmentation
Ultrasound enhancement of fibrinolysis was demonstrated in preclinical models and studied in pilot human stroke trials. Ultrasound can be delivered to an acute cerebral arterial occlusion in several ways, including (1) by a sonographic operator actively positioning a diagnostic Doppler or B-mode/color flow duplex imaging probe285,528,529; (2) by unfocused, low-frequency ultrasound that sonicates both the vessels and brain without imaging guidance291; and (3) intra-arterial or intraclot delivery via catheter, such as with the EKOS technology.532
In the CLOTBUST trial,280 83% of patients achieved any recanalization (46% complete, 27% partial) with intravenous rtPA and TCD versus 50% (17% complete, 33% partial) with intravenous rtPA alone within 2 hours of treatment (P=0.001). The sICH rate was 3.8% in both groups (P=NS).
Because application in humans of frequencies below the diagnostic range resulted in increased symptomatic bleeding rates,291 mechanisms by which megahertz and kilohertz frequencies interact with the clot–residual flow interface and endothelium are currently under renewed investigations, while trials of diagnostic ultrasound enhancement of fibrinolysis are ongoing.531
Combination Intravenous Therapies
Combinations of fibrinolytic(s) plus anticoagulant and/or antiplatelet agents may offer considerable potential to achieve and maintain arterial patency. Multiple exploratory pilot trials have been encouraging, but definitive phase III efficacy trials have yet to be performed.532
As with any medical therapy that carries more than minimal risk, explicit informed patient consent for fibrinolytic therapy is indicated. For the incompetent patient, consent may be provided by a legally authorized representative who can provide proxy consent. A physician’s note documenting explicit discussion in a consent conversation is acceptable. In some institutions, the patient or representative must sign a written consent form conveying the risks and benefits of therapy. In an emergency, when the patient is not competent and there is no available legally authorized representative to provide proxy consent, it is both ethically and legally permissible to proceed with fibrinolysis.533 Generally accepted legal and ethical doctrines recognize an exception to the obligation to obtain explicit informed consent in emergency situations in which immediate treatment is required to prevent more serious harm, the patient lacks decision-making capacity, and no substitute decision maker (surrogate) is available.533–535 Regulatory precedents set by FDA and the Department of Health and Human Services in the United States and by the World Medical Association internationally support the use of intravenous rtPA in patients lacking capacity when an alternative form of consent cannot be obtained within the treatment window.534
Conclusions and Recommendations
Intravenous administration of rtPA remains the only FDA-approved pharmacological therapy for treatment of patients with acute ischemic stroke.11 Its use is associated with improved outcomes for a broad spectrum of patients who can be treated within 3 hours of the last known well time before symptom onset and a mildly more selective spectrum of patients who can be treated between 3 and 4.5 hours of the last known well time. Most importantly, earlier treatment is more likely to result in a favorable outcome. Patients within 3 hours of onset with major strokes (NIHSS score >22) have a very poor prognosis, but some positive treatment effect with intravenous rtPA remains.536 Treatment with intravenous rtPA is associated with increased rates of intracranial hemorrhage, which may be fatal.
Recommendations for the management of intracranial hemorrhage after treatment with intravenous rtPA are provided in the AHA Stroke Council’s updated guideline statement on management of ICH.536a The best methods for preventing bleeding complications are careful selection of patients and scrupulous ancillary care, especially close observation, as well as monitoring of the patient with early treatment of arterial hypertension. Factors that affect decisions about administration of intravenous rtPA are outlined in Tables 10 and 11, and the treatment regimen for administration of intravenous rtPA is included in Table 12. Case series have suggested that fibrinolysis may be used in patients with seizures at the time of presentation when evidence suggests that residual deficits are attributable to ischemia rather than the postictal state.537,538 Additional refinement of relative and absolute contraindications to fibrinolysis needs to be considered. Benefit of therapy has been demonstrated only in trials that avoided concomitant treatment with anticoagulants and antiplatelet agents during the first 24 hours after treatment. Although other fibrinolytic agents, including defibrinogenating drugs, have been tested, none has been established as effective or as a replacement for intravenous rtPA.
Intravenous rtPA (0.9 mg/kg, maximum dose 90 mg) is recommended for selected patients who may be treated within 3 hours of onset of ischemic stroke (Class I; Level of Evidence A). Physicians should review the criteria outlined in Tables 10 and11 (which are modeled on those used in the NINDS Trial) to determine the eligibility of the patient. A recommended regimen for observation and treatment of patients who receive intravenous rtPA is described in Table 12. (Unchanged from the previous guideline13)
In patients eligible for intravenous rtPA, benefit of therapy is time dependent, and treatment should be initiated as quickly as possible. The door-to-needle time (time of bolus administration) should be within 60 minutes from hospital arrival (Class I; Level of Evidence A). (New recommendation)
Intravenous rtPA (0.9 mg/kg, maximum dose 90 mg) is recommended for administration to eligible patients who can be treated in the time period of 3 to 4.5 hours after stroke onset (Class I; Level of Evidence B). The eligibility criteria for treatment in this time period are similar to those for people treated at earlier time periods within 3 hours, with the following additional exclusion criteria: patients >80 years old, those taking oral anticoagulants regardless of INR, those with a baseline NIHSS score >25, those with imaging evidence of ischemic injury involving more than one third of the MCA territory, or those with a history of both stroke and diabetes mellitus. (Revised from the 2009 intravenous rtPA Science Advisory14)
Intravenous rtPA is reasonable in patients whose blood pressure can be lowered safely (to below 185/110 mm Hg) with antihypertensive agents, with the physician assessing the stability of the blood pressure before starting intravenous rtPA (Class I; Level of Evidence B). (Unchanged from the previous guideline13)
In patients undergoing fibrinolytic therapy, physicians should be aware of and prepared to emergently treat potential side effects, including bleeding complications and angioedema that may cause partial airway obstruction (Class I; Level of Evidence B). (Revised from the previous guideline13)
Intravenous rtPA is reasonable in patients with a seizure at the time of onset of stroke if evidence suggests that residual impairments are secondary to stroke and not a postictal phenomenon (Class IIa; Level of Evidence C). (Unchanged from the previous guideline13)
The effectiveness of sonothrombolysis for treatment of patients with acute stroke is not well established (Class IIb; Level of Evidence B). (New recommendation)
The usefulness of intravenous administration of tenecteplase, reteplase, desmoteplase, urokinase, or other fibrinolytic agents and the intravenous administration of ancrod or other defibrinogenating agents is not well established, and they should only be used in the setting of a clinical trial (Class IIb; Level of Evidence B). (Revised from the previous guideline13)
The effectiveness of intravenous treatment with rtPA is not well established (Class IIb; Level of Evidence C) and requires further study for patients who can be treated in the time period of 3 to 4.5 hours after stroke but have 1 or more of the following exclusion criteria: (1) patients >80 years old, (2) those taking oral anticoagulants, even with INR ≤1.7, (3) those with a baseline NIHSS score >25, or (4) those with a history of both stroke and diabetes mellitus. (Revised from the 2009 intravenous rtPA Science Advisory14)
Use of intravenous fibrinolysis in patients with conditions of mild stroke deficits, rapidly improving stroke symptoms, major surgery in the preceding 3 months, and recent myocardial infarction may be considered, and potential increased risk should be weighed against the anticipated benefits (Class IIb; Level of Evidence C). These circumstances require further study. (New recommendation)
The intravenous administration of streptokinase for treatment of stroke is not recommended (Class III; Level of Evidence A). (Revised from the previous guideline13)
The use of intravenous rtPA in patients taking direct thrombin inhibitors or direct factor Xa inhibitors may be harmful and is not recommended unless sensitive laboratory tests such as aPTT, INR, platelet count, and ECT, TT, or appropriate direct factor Xa activity assays are normal, or the patient has not received a dose of these agents for >2 days (assuming normal renal metabolizing function). Similar consideration should be given to patients being considered for intra-arterial rtPA (Class III; Level of Evidence C). (New recommendation) Further study is required.
The number of options for endovascular treatment of ischemic stroke has increased substantially over the past decade to include intra-arterial fibrinolysis, mechanical clot retrieval with the Mechanical Embolus Removal in Cerebral Ischemia (Merci) Retrieval System (Concentric Medical, Inc, Mountain View, CA), mechanical clot aspiration with the Penumbra System (Penumbra, Inc, Alameda, CA), and acute angioplasty and stenting. Intra-arterial fibrinolysis with recombinant prourokinase (r-pro-UK) was studied in the first randomized trial of an endovascular therapy, and this study was published in 1999.168 The Prolyse in Acute Cerebral Thromboembolism (PROACT) II trial of r-pro-UK was positive; however, 2 trials are necessary for any new drug to receive FDA approval. A second trial has not been undertaken, and thus, r-pro-UK has not received FDA approval. Subsequently, the Merci Retrieval System (2004), the Penumbra System (2007), the Solitaire Flow Restoration Device (ev3 Endovascular, Inc, Plymouth, MN; 2012), and the Trevo Retriever (Stryker Neurovascular, Fremont, CA; 2012) were introduced as mechanical means of recanalization based on pivotal studies without noninterventional control groups. None of these devices have an FDA clinical indication because of the need for randomized comparison with medical therapy strategies. However, they were cleared for use by the FDA as mechanical methods for restoring blood flow to occluded arteries based on their comparability to predicate devices; drugs do not have a comparable mechanistic approval pathway. On the basis of FDA clearance of the Merci and Penumbra devices, the Centers for Medicare and Medicaid Services now provides hospital reimbursement for these procedures. There is no approved drug, including alteplase, for intra-arterial use, and therefore, it is not differentially reimbursed compared with intravenous rtPA. It is in this complex regulatory and financial environment that clinical treatment decisions must be made and randomized clinical trials must be conducted.
Evidence for intra-arterial fibrinolysis comes primarily from 2 randomized trials, the randomized PROACT II study and the Middle Cerebral Artery Embolism Local Fibrinolytic Intervention Trial (MELT).168,170 PROACT II was a prospective, phase III randomized trial designed to test the effectiveness of intra-arterial fibrinolysis using r-pro-UK to treat MCA (M1 or M2) occlusions within 6 hours of stroke symptom onset.168 Selection criteria included NIHSS score ≥4 (except isolated aphasia or hemianopia) and age 18 to 85 years. Among the 180 randomized patients, there was an excess of diabetic patients in the control arm (31% versus 13%) and an excess of baseline CT scan protocol violations in the r-pro-UK arm (10% versus 4%). In the primary intention-to-treat analysis, 40% of the 121 patients treated with r-pro-UK and 25% of the 59 control patients had an mRS score of 0 to 2 at 90 days (P=0.04). MCA recanalization was achieved in 66% of the r-pro-UK arm and 18% of the control group (P=0.001). sICH occurred in 10% of patients treated with r-pro-UK and in 2% of the control group (P=0.06). Mortality in the 2 groups was similar.
MELT compared medical management with intra-arterial urokinase within 6 hours and was stopped prematurely because of Japan’s regulatory approval of intravenous rtPA for ischemic strokes within 3 hours.170,539 At stoppage, rates of the primary end point (mRS score 0–2) were numerically higher in the urokinase-treated group than the control group, but this did not reach statistical significance (49.1% versus 36.8%; P=0.35). The preplanned secondary end point (mRS score 0–1) was achieved in 42.1% of urokinase-treated cases and 22.8% of control cases (P=0.045). sICH occurred in 9% of urokinase-treated cases. Both the treatment effect size and sICH rates were consistent with the results of the PROACT II trial, and meta-analysis (combined with PROACT II) showed cumulative evidence in favor of the intra-arterial fibrinolytic approach.540,541
Extrapolation of the randomized trial data to other currently available fibrinolytic agents, including alteplase, is based primarily on consensus and case series data.542–544 The use of intra-arterial fibrinolysis for occlusions in additional locations, such as the basilar artery and intracranial carotid artery, is based primarily on consensus and case series data as well.164,246,545–548 Macleod et al539 randomized 16 patients with angiographic evidence of posterior circulation occlusions who presented within 24 hours of symptom onset to either intra-arterial urokinase or conservative management; both arms underwent anticoagulation with heparin, followed by warfarin. In this small study, good clinical outcomes (defined by a combined mRS and Barthel index end point) were observed in 50% of the intra-arterial urokinase arm compared with 12.5% of the nonurokinase arm (P=0.28).
The intra-arterial approach is thought to be more efficacious for recanalization of proximal arterial occlusions than intravenous fibrinolysis, but the evidence for this is limited. Supportive evidence comes primarily from a cohort study by Mattle et al.245 They compared stroke outcomes at 2 stroke units, each of which treated exclusively with either intravenous rtPA or intra-arterial urokinase. Favorable outcomes (mRS score 0–2) were seen in 29 (53%) of 55 intra-arterial cases and 13 (23%) of 57 intravenous cases (P=0.001). In addition, a small feasibility study by Sen et al549 randomized consecutive patients with proximal arterial occlusions on CTA scan within 3 hours of stroke symptom onset to standard intravenous rtPA (0.9 mg/kg) versus intra-arterial rtPA (up to 22 mg over 2 hours). Median NIHSS scores were 17 and 16 and mean ages were 71 and 66 years for the intravenous and intra-arterial arms, respectively. Fibrinolysis was initiated at a mean of 95 minutes for the intravenous arm and 120 minutes for the intra-arterial arm (P=0.4). The intravenous group had 1 sICH, and the intra-arterial group had 1 asymptomatic ICH. All intra-arterial cases had recanalization, and none of the intravenous cases had recanalization (P=0.03). Neurological improvement (a 4-point decrease in NIHSS score at 90 days) was seen in 3 of 4 patients treated with intravenous rtPA and 2 of 3 treated with intra-arterial rtPA.
On the basis of the premise that intra-arterial therapy may be more effective for recanalization of larger thrombi, severe neurological deficits (NIHSS score ≥10) that suggest a proximal arterial occlusion and radiographic evidence of occlusion of a major intracranial vessel have been considered potential indications for the use of intra-arterial therapy. However, this clinical benefit may be counterbalanced by delay to treatment initiation with the intra-arterial approach and consequent late reperfusion, potential risks of periprocedural sedation, and treatment-related complications. Definitive data from randomized controlled trials delineating the relative efficacy of intra-arterial therapy versus intravenous rtPA treatment are lacking at this time.
Intra-arterial fibrinolysis is a consideration for patients ineligible for intravenous rtPA. For example, the PROACT II trial may be applicable to patients eligible for treatment within 6 hours; more definitive data for patients in the extended time window from randomized controlled trials are needed.550 Recent history of a major surgical procedure poses systemic bleeding risk in the setting of intravenous rtPA and may represent another group for consideration of intra-arterial fibrinolysis. Several small case series of postoperative cardiac surgery cases suggest reasonable safety of intra-arterial fibrinolysis.551–553 In addition, a retrospective case series of 36 ischemic stroke patients from 6 academic centers treated with intra-arterial fibrinolysis after surgical procedures, including open heart surgery (n=18), CEA (n=6), and urologic-gynecologic surgery (n=4), suggested that intra-arterial rtPA is reasonably safe in the postoperative setting, with the exception of neurosurgical procedures (n=3).554 Major systemic bleeding occurred in 4 cases, including 3 postcraniotomy ICHs and 1 post–coronary artery bypass graft hemopericardium.
Rates of good clinical outcome after intra-arterial fibrinolysis are likely to be highly time dependent, as is the case with intravenous rtPA treatment.92,93,555 If intra-arterial fibrinolysis treatment is planned, an emphasis should be placed on rapid triage, patient transport, and clinical team mobilization.
Combination Intravenous and Intra-arterial Fibrinolysis
Initial studies of fibrinolytic therapy in acute ischemic stroke involved a single pharmacological agent, alteplase, given either intravenously or intra-arterially. It was subsequently proposed that combined intravenous and intra-arterial fibrinolysis may be a more efficient way to rapidly recanalize major intracranial arterial occlusions. This would allow for immediate initiation of intravenous fibrinolysis in an ED, followed by rapid mobilization of the neuroangiographic team and transport of the patient to the angiographic suite for further titrated intra-arterial fibrinolytic therapy, if necessary. This approach could address the concern that delays to intra-arterial therapy may negate the potential benefits of more efficacious recanalization. Proximal intracranial arterial occlusions (distal internal carotid artery, MCA, or basilar artery) may benefit most from this approach because of larger clot burdens that would be more likely to fail treatment with intravenous rtPA alone.
A series of pilot trials have evaluated the combined intravenous/intra-arterial fibrinolytic approach using low-dose rtPA.556–558 The Emergency Management of Stroke Bridging trial was a retrospective analysis of 20 patients with severe stroke who received intravenous and intra-arterial rtPA within 3 hours from symptom onset.558 Despite a median baseline NIHSS score of 21, 50% of patients recovered to an mRS score of 0 to 1 on follow-up. The feasibility and suggestion of efficacy led to the creation of the Interventional Management of Stroke (IMS) study. The IMS study enrolled 80 patients 18 to 80 years old with an initial NIHSS score ≥10 who presented within 3 hours of stroke onset.556 Patients received intravenous rtPA (0.6 mg/kg, 60 mg maximum over 30 minutes) started within 3 hours of stroke symptom onset, followed by additional intra-arterial rtPA (up to 22 mg) at the site of the thrombus if there was a persistent occlusion. The median baseline NIHSS score was 18. The rate of sICH (6.3%) was similar to that of comparable intravenous rtPA–treated subjects (6.6%) in the NINDS rtPA Stroke Trial. The 3-month mortality rate (16%) was similar to the placebo (24%) and intravenous rtPA (21%) arms of the NINDS rtPA Stroke Trial. Reperfusion, as quantified by the Thrombolysis in Cerebral Infarction (TICI) score, which attempts to standardized flow restoration reporting in clinical trials559 (TICI score 2–3 indicates good reperfusion), was seen in 56% of cases. Good clinical outcomes (mRS score 0–2) were seen in 43% of cases. The subsequent IMS II study enrolled 81 additional patients and, together with combined intravenous/intra-arterial rtPA, delivered low-energy ultrasound by use of the EKOS system whenever possible. The sICH rate (9.9%) and mortality rate (16%) were again comparable to the NINDS rtPA trial. Reperfusion (TICI score 2–3) was seen in 61% of cases. Good clinical outcomes (mRS score 0–2) were seen in 46% of cases. Both studies showed better outcomes than comparable NINDS placebo cases, and IMS II showed statistically better outcomes in secondary outcome analyses. The phase III IMS III trial, with a planned enrollment of 900 patients with an NIHSS score ≥10 treated within 3 hours of stroke symptom onset, was recently stopped for reported futility; further results from the study are pending.560
Shaltoni et al561 evaluated the combined approach using full-dose (0.9 mg/kg) rtPA followed by intra-arterial fibrinolysis (with reteplase, alteplase, or urokinase) in a prospective cohort of ischemic stroke patients at a single center who presented within 3 hours of symptom onset. These patients were routinely offered intra-arterial therapy if they had a persisting disabling neurological deficit or a persistent or reoccluding thrombus by TCD after they completed the 60-minute intravenous rtPA infusion. The sICH rate was 5.8% (4/69) and the mortality rate was 17.4% (12/69). Partial or complete reperfusion (TICI score 2–3) was seen in 72.5% of cases, and favorable outcome (discharge to acute rehabilitation or home) was seen in 55% of cases.
As with intravenous fibrinolysis, reducing the time to reperfusion with endovascular therapies is likely pivotal in achieving the best clinical outcomes. This is supported by a post hoc pooled analysis of the IMS I and II pilot trials that showed that time to reperfusion, as estimated by the time from symptom onset to completion of the intra-arterial procedure, was an independent predictor of the probability of good clinical outcome. When the time to reperfusion was increased by 30 minutes, from 280 to 310 minutes, the probability of a favorable outcome (mRS score 0–2) was 10.6% less likely.555
Mechanical Clot Disruption/Extraction
Mechanical thrombectomy is a consideration as both a primary reperfusion strategy and in conjunction with pharmacological fibrinolysis for achieving recanalization in patients with acute ischemic stroke.562 Recanalization by this means may occur because of a combination of thrombus fragmentation, thrombus retrieval, and enhancement of fibrinolytic penetration. There are currently 4 devices cleared by the FDA for recanalization of arterial occlusion in patients with ischemic stroke. The Merci Retrieval System received FDA clearance in 2004 and consists of the Merci Retriever, the Merci Balloon Guide Catheter, and the Merci Microcatheter. The Merci Retriever uses a memory-shaped nitinol wire with helical loops of decreasing diameter at its distal end to engage the clot. It is advanced through the microcatheter in its compressed form distal to the occlusion. Subsequent withdrawal of the microcatheter deploys the device in its preimposed helical shape. Since initial FDA clearance, the retriever design has been updated, with the newest V series retriever having a series of loops to engage and capture the clot. The Penumbra System received FDA clearance in 2007 and consists of the aspiration pump, reperfusion catheters, and separators. It is designed to aspirate thrombus from large intracranial vessels by placing a reperfusion catheter at the proximal end of the thrombus and connecting it to a vacuum source. A continuous aspiration-debulking process is facilitated by advancing and withdrawing the separator through the Penumbra reperfusion catheter. Since initial FDA clearance, the reperfusion catheter has been modified with a larger, tapered lumen and new polymer composition at the distal end to increase accessibility and aspiration efficiency. A further update consisting of a 3-dimensional separator is under investigational study. Most recently, the Solitaire Flow Restoration Device and the Trevo Retriever received FDA clearance in 2012. These are both retrievable stents that are deployed within the thrombus to displace it radially, incorporate it within the stent’s struts, and then extract it.
The Merci Retriever was evaluated in patients ineligible for intravenous rtPA and with arterial occlusions who presented within 8 hours of stroke symptom onset in the pivotal single-arm, prospective, multicenter MERCI trial.563 Recanalization was achieved in 46% (n=69) of the 151 patients on intention-to-treat analysis and in 48% (n=68) of the 141 patients in whom the device was deployed. Clinically significant procedural complications and sICH occurred in 7% and 8% of the patients, respectively. Good neurological outcomes (mRS score 0–2) at 90 days were observed more frequently in patients with successful recanalization than in those with unsuccessful recanalization (46% versus 10%, P<0.0001). The Multi MERCI trial564 studied thrombectomy in patients with ischemic stroke and large-vessel occlusion treated within 8 hours of symptom onset with newer-generation retriever devices. Patients with persistent occlusion after intravenous rtPA treatment were included. One hundred sixty-four patients were treated with thrombectomy, and 131 were treated initially with the new-generation retrievers. Treatment with the new-generation retriever resulted in successful recanalization in 57% of treated arteries and in 70% after adjunctive therapy (intra-arterial fibrinolysis or other mechanical devices). Overall, favorable clinical outcomes (mRS score 0–2) were seen in 36% of the patients, and 34% of the patients died. Clinically significant procedural complications and sICH occurred in 6% and 10% of the patients, respectively.
A subgroup analysis of Multi MERCI trial compared outcomes between patients who did or did not receive intravenous rtPA before thrombectomy.565 Thirty patients (27%) received intravenous rtPA before thrombectomy. The sICH rate was 7% and 10% in patients pretreated and not pretreated with intravenous rtPA, respectively. Two subgroup analyses compared outcomes in patients with arterial occlusion located at particular sites in the MERCI and Multi MERCI trials. Of the 80 patients with intracranial internal carotid artery occlusion,566 53% and 63% had recanalization with the retriever alone and with the retriever and additional endovascular treatment, respectively. Good clinical outcome (mRS score 0–2) at 90 days occurred in 39% of patients with recanalization and in 3% of patients without recanalization. Recanalization remained a significant predictor of a good 90-day outcome in multivariate analysis. In another analysis of 27 patients with vertebrobasilar arterial occlusions, recanalization occurred in 78% of patients after retriever use in the MERCI and Multi MERCI trials.567 Good clinical outcome (mRS score 0–3) at 90 days occurred in 41% of patients, and 44% died. Another analysis of patients recruited in the MERCI and Multi MERCI trials compared outcomes between patients with abnormal INR >1.7, PTT >45 seconds, or platelet count <100 000/µL and those with normal hemostasis.568 Rates of partial or complete recanalization, mortality, or major sICH were not significantly different; however, the rate of favorable outcomes was substantially lower among those with abnormal hemostasis (9% versus 35%, P=0.002). Another subgroup analysis compared outcomes in similar patients from the MERCI and Multi MERCI cohorts with historical comparators from the active and control arms of the PROACT II trial. Mechanical thrombectomy produced rates of good clinical outcomes (mRS score 0–2; 39.9%) similar to PROACT II patients treated with intra-arterial pro-UK (39.5%) compared with PROACT II control patients (25.4%).569
The pivotal Penumbra trial was a prospective, multicenter, single-arm study570 of 125 patients with NIHSS scores ≥8 who presented within 8 hours of symptom onset and were treated with the Penumbra System.570 Patients who presented within 3 hours from symptom onset were either ineligible for intravenous rtPA or refractory to intravenous rtPA. Partial or complete recanalization was reported in 82% of the treated vessels, although the operational method for characterizing recanalization was not specified. Procedural complications and sICH occurred in 13% and 11% of the patients, respectively. Overall, favorable clinical outcomes (mRS score 0–2) were seen in 25% of the patients, and 33% of the patients died. Subsequently, Tarr and colleagues571 conducted a post–FDA approval multicenter retrospective case review of 157 consecutive patients treated with the Penumbra system. Partial or complete target-vessel recanalization was achieved in 87% of patients (54% with Thrombolysis in Myocardial Infarction [TIMI] grade 2 and 33% with TIMI grade 3). Procedural events occurred in 9 patients and device malfunctions in 3. sICH, defined by any evidence of ICH on CT within 24 hours after the procedure and a deterioration of the NIHSS score by >4 points, occurred in 6.4% of patients. At 90 days after stroke, 41% of patients had achieved an mRS score of 0 to 2, and all-cause mortality was 20%.
The pivotal studies of the Solitaire and Trevo devices were published most recently.572,573 The SWIFT study (Solitaire FR With the Intention for Thrombectomy) compared the recanalization efficacy of Solitaire with the Merci Retrieval System in a randomized, prospective noninferiority trial of 113 subjects with moderate or severe strokes. Eligible subjects were within 8 hours of symptom onset and were either ineligible for or refractory to intravenous rtPA. After a prespecified interim analysis led to early halting of the trial, successful revascularization (TIMI 2–3 recanalization) without symptomatic intracranial hemorrhage was reported in 61% of Solitaire cases versus 24% of the MERCI group (P<0.001) based on a blinded assessment. This corresponded to 90-day good neurological outcome rates (mRS score 0–2) of 58% versus 33% (P=0.001), respectively, and 90-day mortality rates of 17% versus 38% (P=0.001), respectively. The TREVO 2 study (Thrombectomy REvascularization of large Vessel Occlusions) was a similar design with the exception of the primary outcome definition. In this case, the Trevo Retriever was compared with the Merci Retriever in a randomized noninferiority study of 178 subjects. The primary outcome was TICI 2 to 3 angiographic reperfusion assessed in an unblinded manner. The study reported revascularization rates of 86% in the Trevo group versus 60% in the MERCI group (P<0.0001). Correspondingly, 90-day good clinical outcomes (mRS score 0–2) were seen in 40% versus 22%, respectively (P=0.01), and 90-day mortality was seen in 33% versus 24%, respectively (P=0.18). Both studies supported superiority of their devices compared with the predicate Merci device and concluded that prospective randomized studies compared with medical treatment alone were needed.
The IMS III trial studied intravenous rtPA alone compared with combined intravenous rtPA and endovascular therapies including mechanical devices (largely Merci and Penumbra) as an option for the combined intravenous/intra-arterial approach being tested, with the hope of providing additional safety and efficacy data for this approach. It was halted early on the basis of a prespecified interim analysis that demonstrated futility, and detailed results are pending.574
Acute Angioplasty and Stenting
Intracranial Acute Angioplasty and Stenting
Increasingly, urgent angioplasty with adjunctive stent deployment is being used to restore antegrade flow, with or without fibrinolysis or clot extraction. The nonrandomized, single-center Stent-Assisted Recanalization in Acute Ischemic Stroke (SARIS) study suggested that direct stenting of the occluded culprit vessel, at least for intracranial locations, is technically effective in restoring flow promptly.575 Among 20 patients ineligible for or not responsive to intravenous rtPA, partial or complete recanalization was achieved in all patients, sICH occurred in 5%, and fair or better functional outcomes (mRS score 0–3) at 1 month were seen in 60%. The SARIS study provides evidence that additional patients with acute stroke might benefit from expeditious reperfusion with stents, but this approach requires additional study.
Retrievable stents are the newest approach to endovascular recanalization. Examples include the Solitaire FR and Trevo devices. These stent retrievers are deployed within symptomatic intracranial thrombi to reperfuse tissue immediately and then used to engage and retrieve the clot. Removal of the stent eliminates the need for acute double-antiplatelet therapy, as is needed for permanent stent placement. Current data, which are limited to case series, suggest high (80%–90%) recanalization rates and reasonable safety.576,577 Registries and additional randomized controlled studies are also under way.
Extracranial Acute Angioplasty and Stenting
Angioplasty and stenting of extracranial carotid (or extracranial vertebral arteries) is predominantly performed for stroke prevention rather than acute stroke treatment. However, this therapy has been used on an emergency basis in the setting of acute stroke for 2 situations in particular: when the primary cause of the stroke is attenuation or cessation of flow in the extracranial carotid or vertebral artery, such as with total or near-total occlusion caused by severe atherosclerosis or dissection, and when catheter access to a culprit intracranial thrombus is impeded by severe stenosis of the extracranial carotid, and angioplasty/stenting of the carotid is required before treatment of a more distal intracranial occlusion.
Although there are no completed prospective, randomized controlled trials demonstrating relative efficacy and safety of angioplasty and stenting of the extracranial carotid in acute ischemic stroke, small retrospective case series have reported promising results.578–585 Nedeltchev et al582 described angioplasty and stenting of the internal carotid artery in conjunction with intra-arterial fibrinolysis in 25 patients who had acute carotid artery occlusion that caused MCA territory ischemic stroke and compared them with a group of 31 medically treated patients. Favorable outcomes were more frequent (56% versus 26%) among patients who received endovascular treatment. Jovin et al581 showed that emergency revascularization of internal carotid occlusion with a carotid stent had a high success rate (23 of 25 patients) with low rates of adverse events. Similarly, Nikas et al578 showed a high rate of procedural success (83%) in 14 patients with atheromatous obstruction and 4 patients with dissection of the internal carotid artery. Imai et al580 demonstrated that an emergency carotid stent can improve 7-day neurological outcome and may improve midterm clinical outcomes compared with historical controls. In selected patients with acute vertebrobasilar ischemic stroke, angioplasty and stenting of the vertebral artery has been combined with emergency administration of fibrinolytic agents.585
The relative role of endovascular versus surgical revascularization of the extracranial carotid artery emergently in acute stroke remains to be determined. No studies have yet been performed to compare the utility of these alternative approaches for revascularization of the extracranial internal carotid artery in acute stroke. Additional studies must be undertaken to define the role of angioplasty and stenting of the extracranial carotid arteries in the early management of acute stroke.
More emphasis has been placed on deriving information from the initial and postrevascularization angiograms, with emphasis on the site of occlusion, identification of collateral supply to the affected region, and precise definitions of revascularization. There are new data that suggest that this information may be incorporated into a scheme to stratify patients with regard to expected rate of recanalization and short-term outcome after intra-arterial fibrinolysis. The angiographic results of cerebral reperfusion procedures were initially characterized with the TIMI grading system, a 4-point scale from 0 (complete occlusion) to 3 (complete reperfusion) that was originally developed to assess arterial occlusion and perfusion in patients with myocardial infarction.586 However, the TIMI grading system has several limitations. It does not account for occlusion location or collateral circulation. Even as a measure of anterograde reperfusion, the cardiac TIMI scale cannot be applied to the more complex cerebral vasculature without the creation of additional operational rules. Under the rubric “TIMI scale,” recent stroke clinical trials have actually used very different brain-adapted versions of the TIMI, which hampers comparisons and understanding of trial findings.587 The Qureshi grading system is a scale from 0 (best possible score) to 5 (worst possible score) that angiographically classifies location of arterial occlusions before and after recanalization.588–590 Other studies have placed emphasis on 2 scales developed specifically for the cerebral circulation to measure recanalization of the primary arterial occlusive lesion and global reperfusion of the distal vascular bed.530,591 The Arterial Occlusive Lesion (AOL) score is defined on a scale of 0 to 3, ranging from no recanalization to complete recanalization of the primary occlusion. The TICI score was developed in 2003 in an effort to standardize reporting of revascularization efforts. The TICI score is defined from 0 to 3, ranging from no perfusion to full perfusion with filling of all distal branches.559 TICI is currently being used in the IMS trial560 and an ongoing stroke registry.592
Additional studies have examined reocclusion and distal fragmentation after a combination of pharmacological fibrinolysis and mechanical thrombectomy. In an analysis of data from 4 prospective acute stroke protocols,593 distal embolization was defined qualitatively as appearance of an occlusion on a downstream vessel, and arterial reocclusion was defined as subsequent reocclusion of the target vessel after initial recanalization had been achieved. Arterial reocclusion occurred in 18% of these patients, whereas distal embolization occurred in 16% of the 91 patients treated in these protocols. Arterial reocclusion, but not distal embolization, was associated with a lower likelihood of favorable outcome at 1 to 3 months after adjustment for potential confounders. Another analysis of 56 patients594 who underwent cerebral angiography at 24 hours to determine the status of occlusion after endovascular treatment (compared with immediate postprocedure angiogram) observed subacute recanalization in 16 patients (29%), including additional recanalization in 8 patients with early recanalization. Subacute reocclusion was observed in 5 patients (9%). Subacute recanalization was associated with a trend toward a higher rate of favorable outcome after adjustment for other covariates.
Conclusions and Recommendations
A number of techniques and devices are under study in several trials. Although several devices have resulted in recanalization with acceptable safety, direct comparative data between the devices are not available. The combination of pharmacological fibrinolysis and mechanical thrombectomy appears to have the highest rate of recanalization without any difference in rate of intracranial hemorrhage. As the rate of recanalization has increased, new challenges such as reocclusion, distal fragmentation, and lack of clinical benefit despite complete recanalization have been identified. Consistently, recanalization rates in trials exceed rates of the best clinical outcomes, which suggests the importance of patient selection independent of the technical effectiveness of thrombectomy devices. As with the intra-arterial administration of fibrinolytics, the use of these devices will be limited to those CSCs that have the resources and physician expertise to perform these procedures safely.595 Lastly, as with intravenous fibrinolysis, time is brain for all forms of endovascular reperfusion, and all efforts must be made to reduce time to reperfusion, because the likelihood of favorable outcome is directly linked to the time to reperfusion.555
Patients eligible for intravenous rtPA should receive intravenous rtPA even if intra-arterial treatments are being considered (Class I; Level of Evidence A). (Unchanged from the previous guideline13)
Intra-arterial fibrinolysis is beneficial for treatment of carefully selected patients with major ischemic strokes of <6 hours’ duration caused by occlusions of the MCA who are not otherwise candidates for intravenous rtPA (Class I; Level of Evidence B). The optimal dose of intra-arterial rtPA is not well established, and rtPA does not have FDA approval for intra-arterial use. (Revised from the previous guideline13)
As with intravenous fibrinolytic therapy, reduced time from symptom onset to reperfusion with intra-arterial therapies is highly correlated with better clinical outcomes, and all efforts must be undertaken to minimize delays to definitive therapy (Class I; Level of Evidence B). (New recommendation)
Intra-arterial treatment requires the patient to be at an experienced stroke center with rapid access to cerebral angiography and qualified interventionalists. An emphasis on expeditious assessment and treatment should be made. Facilities are encouraged to define criteria that can be used to credential individuals who can perform intra-arterial revascularization procedures. Outcomes on all patients should be tracked (Class I; Level of Evidence C). (Revised from the previous guideline13)
When mechanical thrombectomy is pursued, stent retrievers such as Solitaire FR and Trevo are generally preferred to coil retrievers such as Merci (Class I; Level of Evidence A). The relative effectiveness of the Penumbra System versus stent retrievers is not yet characterized. (New recommendation)
The Merci, Penumbra System, Solitaire FR, and Trevo thrombectomy devices can be useful in achieving recanalization alone or in combination with pharmacological fibrinolysis in carefully selected patients (Class IIa; Level of Evidence B). Their ability to improve patient outcomes has not yet been established. These devices should continue to be studied in randomized controlled trials to determine the efficacy of such treatments in improving patient outcomes. (Revised from the previous guideline13)
Intra-arterial fibrinolysis or mechanical thrombectomy is reasonable in patients who have contraindications to the use of intravenous fibrinolysis (Class IIa; Level of Evidence C). (Revised from the previous guideline13)
Rescue intra-arterial fibrinolysis or mechanical thrombectomy may be reasonable approaches to recanalization in patients with large-artery occlusion who have not responded to intravenous fibrinolysis. Additional randomized trial data are needed (Class IIb; Level of Evidence B). (New recommendation)
The usefulness of mechanical thrombectomy devices other than the Merci retriever, the Penumbra System, Solitaire FR, and Trevo is not well established (Class IIb; Level of Evidence C). These devices should be used in the setting of clinical trials. (Revised from the previous guideline13)
The usefulness of emergent intracranial angioplasty and/or stenting is not well established. These procedures should be used in the setting of clinical trials (Class IIb; Level of Evidence C). (New recommendation)
The usefulness of emergent angioplasty and/or stenting of the extracranial carotid or vertebral arteries in unselected patients is not well established (Class IIb; Level of Evidence C). Use of these techniques may be considered in certain circumstances, such as in the treatment of acute ischemic stroke resulting from cervical atherosclerosis or dissection (Class IIb; Level of Evidence C). Additional randomized trial data are needed. (New recommendation)
For >50 years, physicians have prescribed intravenously administered anticoagulants for treatment of patients with acute ischemic stroke, but these medications are now used less often.596,597 The cited reasons for emergency use of these medications to treat stroke include (1) to halt neurological worsening, (2) to prevent early recurrent embolization, and (3) to improve neurological outcomes. Past panels of the AHA concluded that the data about the safety and efficacy of heparin or other emergently administered anticoagulants were either negative or inconclusive.11,13,598,599 Other groups also have concluded that the data from clinical trials have not established the utility of emergency anticoagulation in treatment of patients with recent ischemic stroke.143,600,601
Anticoagulants often were prescribed to patients with recent stroke in an effort to prevent early recurrent cardioembolic stroke, including those with atrial fibrillation. The Cerebral Embolism Study Group estimated that the risk of early recurrent embolism was ≈12% among untreated patients with embolic stroke.602,603 Subsequently, a trial found that the risk of recurrent stroke within 1 week was ≈8% among patients with atrial fibrillation.604 Other trials testing anticoagulants administered immediately after stroke have reported much lower rates (≈0.3%–0.5% per day).605–607 These relatively low rates mean that detection of a therapeutic effect from anticoagulants for prevention of early recurrent embolism will be difficult to achieve.
The International Stroke Trial (IST) tested subcutaneously administered unfractionated heparin (UFH) in doses of 5000 or 25 000 U/d started within 48 hours of stroke.606 Dual randomization meant that approximately half of the patients receiving heparin were also prescribed aspirin. Neither monitoring of the level of anticoagulation nor adjustment of dosages in response to levels of anticoagulation was performed. In addition, some patients did not have a brain imaging study before entry into the trial, and thus, some patients with hemorrhagic stroke may have been enrolled. Although heparin was effective in lowering the risk of early recurrent stroke, an increased rate of bleeding complications negated this benefit. A subgroup analysis did not find a benefit from heparin in lowering the risk of recurrent stroke among those patients with atrial fibrillation.608
Other studies of anticoagulation similarly failed to show definitive benefit. A Swedish study failed to demonstrate a benefit from heparin for treatment of patients with progressing stroke.609 A single-center Italian trial enrolled patients within 3 hours after onset of stroke and treated patients with an infusion of intravenous heparin starting with a bolus dose, with adjustments in dosage in response to aPTT.610 Thirteen of 208 heparin-treated patients had symptomatic hemorrhagic complications (6.2%; 7 fatal), whereas 3 of 210 control patients (1.4%) had sICH. Favorable outcomes at 90 days were reported in 81 patients treated with heparin (38.9%) and 60 control patients (28.6%). Given the results of this trial, the authors concluded that additional study of very early administration of heparin in patients with cardioembolic stroke was reasonable.611 A multicenter European trial administered heparin to 32 patients and aspirin to 35 patients before it was halted prematurely.612 The investigators reported no significant differences in outcomes, rates of recurrent ischemic stroke, symptomatic hemorrhage, or death between the 2 treatment groups. Sandercock et al613 performed a systemic review of anticoagulants in treatment of acute ischemic stroke and concluded that treatment with immediate anticoagulant therapy was not associated with any net short- or long-term benefit.
A meta-analysis of anticoagulants in patients with presumed cardioembolic stroke found that the agents were associated with a nonsignificant reduction in the rate of early recurrent stroke, an increased risk of ICH, and no reduction in either death or disability.614 The safety and efficacy of heparin, given as an interim therapy for those patients with atrial fibrillation who were beginning to receive oral anticoagulants, was evaluated in an observational study.615 Heparin did not reduce the risk of thromboembolic events or increase the risk of bleeding complications, but the heparin bridging did prolong hospitalization. Besides an associated risk of bleeding, the administration of heparin to patients with acute ischemic stroke may be complicated by the development of heparin-induced thrombocytopenia.616
Lower-Molecular-Weight Heparins and Danaparoid
The utility of several different low-molecular-weight heparins (LMWHs) or danaparoid in treating patients with acute ischemic stroke has been evaluated in clinical trials. Most trials tested subcutaneous administration of these anticoagulants. Some trials compared these medications to UFH or aspirin, whereas others have compared these medications to control or placebo. Generally, the results of these trials were negative.
Early increased hemorrhage risk was found in most early LMWH trials, outweighing early prevention benefits. Kay et al617 tested 2 doses of nadroparin given over a 10-day period after stroke. Although a benefit was not found at 3 months, those who received the larger dose of nadroparin had a significantly lower mortality at 6 months than the control group. Another trial of nadroparin did not find improvement in favorable outcomes but found an increased risk of bleeding with the higher of the 2 doses of the medication.618 In a Norwegian trial, dalteparin was not more effective than aspirin in preventing recurrent events, and more bleeding was seen with the LMWH.604 A subsequent subgroup analysis did not demonstrate any group of patients who would have benefited from dalteparin.619 Similar trials of certoparin and tinzaparin demonstrated no differences in the rates of favorable outcomes.620,621
Intravenous administration of danaparoid (heparinoid/ORG 10172) using a bolus to initiate therapy was tested in a randomized, double-blind, placebo-controlled trial.607 The trial halted recruitment of patients with moderately severe stroke (NIHSS scores >15) because of an increased risk of symptomatic hemorrhage. Danaparoid did not lessen the risk of early recurrent stroke or neurological worsening or improve outcomes at 3 months. The trial included prespecified subgroup analyses among patients with different subtypes of ischemic stroke. The only subgroup that showed potential benefit from treatment was those subjects who had stroke secondary to large-artery atherosclerosis (>50% stenosis), in which favorable outcomes were noted in 64 of 119 patients treated with danaparoid (53.8%) and 41 of 108 patients given placebo (38.0%; P=0.023) at 7 days.622 This finding is supported by the results of a study that found that the likelihood of early recurrent stroke was greatest among people with severe atherosclerotic disease of large arteries.623 As a result, a randomized trial in Singapore and Hong Kong compared aspirin or nadroparin administered within 48 hours of stroke to Asian patients with occlusive disease of larger arteries.624 Almost all of the patients had severe stenosis or occlusions of intracranial arteries, but the trial enrolled few patients with extracranial disease. No differences in the rate of hemorrhage or rates of favorable outcomes were found. Woessner et al625 studied the usefulness of subcutaneously administered enoxaparin or adjusted-dose heparin in a multicenter trial that enrolled patients with either high-grade arterial stenoses or a cardioembolic source; no significant differences were noted between the 2 groups.
Bath et al626 performed a meta-analysis of trials that tested aspirin or LWMHs. They found that the LMWHs significantly reduced the risk of venous thromboembolism but increased the risk of symptomatic bleeding. No differences were found in mortality, rate of recurrent stroke, or rate of neurological worsening. They concluded that LMWH should not replace aspirin in the routine management of patients with ischemic stroke. Another trial compared enoxaparin or UFH for prevention of thromboembolic events among patients with stroke that caused lower-limb paralysis; the 2 medications were equally effective.627 Diener et al628 compared certoparin or heparin in prevention of thromboembolic events after stroke. The LMWH was found to be at least as effective as UFH for prevention of these complications. In the Prevention of VTE After Acute Ischemic Stroke With LMWH Enoxaparin (PREVAIL) study, the usefulness of subcutaneous administration of either heparin or enoxaparin was tested for the prevention of symptomatic or asymptomatic DVT or pulmonary embolism (PE).629 The risk of venous thromboembolism was significantly less with enoxaparin (68 [10%] versus 121 [18%]; risk ratio, 0.57; 95% CI, 0.44–0.76; P=0.001.) The rates of hemorrhage were similar in the 2 treatment groups. Overall, this study gives the strongest evidence of the superiority of LMWH in prevention of venous thromboembolism after ischemic stroke. In 2008, Sandercock et al630 published an update of the Cochrane Systemic Review comparing the utility of UFH and LMWH. They found that the LMWHs were effective in lowering the risk of DVT, but the data were insufficient to determine whether these medications were superior to UFH when other potential therapeutic end points were examined.
Anticoagulants as an Adjunctive Therapy
The administration of either antiplatelet agents or anticoagulants is currently contraindicated during the first 24 hours after treatment with intravenous rtPA. The restriction is based on the clinical trial protocol used in the NINDS trials.166 However, arterial reocclusion may follow successful recanalization with fibrinolysis.290,593,594 In addition, cardiologists often prescribe anticoagulants and antiplatelet agents as part of a multimodality treatment regimen for management of acute coronary artery occlusions. Thus, there is interest in the use of an anticoagulant or antiplatelet agent that may maintain arterial patency after fibrinolytic therapy. The trials of intra-arterially administered r-pro-UK used heparin as part of the treatment regimen, and the control group received only heparin.168,631,632 In the first study, both the success of recanalization and the risk of bleeding were increased among the subjects who received the larger of the 2 doses of adjunctive heparin. Intravenous heparin has been administered after administration of intravenous rtPA.633,634 No increase in bleeding complications was reported. Heparin has been given in addition to abciximab with a reasonable degree of safety635; however, neither the safety nor efficacy of adjunctive anticoagulation has been established, and additional research is required.
Direct thrombin inhibitors may be useful in acute ischemic stroke because of their actions that limit thrombosis. These medications could be considered as an alternative to anticoagulants, and they could be administered to those people who develop heparin-associated thrombocytopenia. Dabigatran, a direct thrombin inhibitor, has been evaluated over the past decade for the prevention of thromboembolic events in patients after orthopedic procedures. More recently, in the RE-LY study (Randomized Evaluation of Long-term Anticoagulation Therapy), dabigatran demonstrated benefit compared with warfarin for the prevention of stroke or systemic embolization in patients with atrial fibrillation.636 At lower doses, dabigatran was noninferior to warfarin while demonstrating fewer hemorrhagic complications. At higher doses, dabigatran was more effective than warfarin but had similar bleeding risk. In October 2010, the FDA approved the higher 150-mg twice-a-day dose for stroke prevention in patients with atrial fibrillation. For patients with impaired renal function, a lower 75-mg twice-a-day dose is recommended. A dose-escalation study of argatroban, also a direct thrombin inhibitor, found that it prolonged aPTT levels but did not increase mortality or the risk of serious bleeding.637 A Japanese study retrospectively examined the impact of argatroban on outcomes among patients with cardioembolic stroke.638 It concluded that argatroban may be superior to heparin in reducing mortality and improving outcomes after strokes. A single case in which argatroban was successfully administered in addition to intravenous and intra-arterial fibrinolysis was also reported.637 Additional research is ongoing regarding the role of argatroban in the treatment of patients with acute stroke.
Conclusions and Recommendations
The results of several clinical trials demonstrate there is an increased risk of bleeding complications with early administration of either UFH or LMWH. Early administration of anticoagulants does not lessen the risk of early neurological worsening. Data indicate that early administration of UFH or LMWH does not lower the risk of early recurrent stroke, including among people with cardioembolic sources. Data are insufficient to indicate whether anticoagulants might be effective among some potentially high-risk groups, such as those people with intracardiac or intra-arterial thrombi. The effectiveness of urgent anticoagulation is not established for treatment of patients with arterial dissection or vertebrobasilar disease. The role of anticoagulants as an adjunct in addition to mechanical or pharmacological fibrinolysis has not been established.
Dabigatran was recently approved for the prevention of stroke and systemic embolism in patients with atrial fibrillation. The timing of initiation after stroke and the usefulness of other antithrombin medications have not been established.
At present, the usefulness of argatroban or other thrombin inhibitors for treatment of patients with acute ischemic stroke is not well established (Class IIb; Level of Evidence B). These agents should be used in the setting of clinical trials. (New recommendation)
The usefulness of urgent anticoagulation in patients with severe stenosis of an internal carotid artery ipsilateral to an ischemic stroke is not well established (Class IIb; Level of Evidence B). (New recommendation)
Urgent anticoagulation, with the goal of preventing early recurrent stroke, halting neurological worsening, or improving outcomes after acute ischemic stroke, is not recommended for treatment of patients with acute ischemic stroke (Class III; Level of Evidence A). (Unchanged from the previous guideline13)
Urgent anticoagulation for the management of noncerebrovascular conditions is not recommended for patients with moderate-to-severe strokes because of an increased risk of serious intracranial hemorrhagic complications (Class III; Level of Evidence A). (Unchanged from the previous guideline13)
5. Initiation of anticoagulant therapy within 24 hours of treatment with intravenous rtPA is not recommended (Class III; Level of Evidence B). (Unchanged from the previous guideline13)
Aspirin is the antiplatelet agent that has been tested the most extensively. Two large trials each demonstrated a nonsignificant trend in reduction in death or disability when treatment with aspirin was begun within 48 hours of stroke.605,606 A minor increase in bleeding complications was found. When the data from the 2 trials were combined, a modest but statistically significant benefit was noted with aspirin therapy. The primary effect was likely attributable to prevention of recurrent events. It is not clear whether aspirin limited the neurological consequences of the acute stroke itself.
There has been limited experience with the use of clopidogrel or dipyridamole in the setting of acute stroke. Initiation of treatment with clopidogrel in a daily dose of 75 mg does not produce maximal inhibition of platelet aggregation for ≈5 days.639 This delay presents a problem for an early treatment effect in the management of patients with acute ischemic stroke. A bolus dose of 300 to 600 mg of clopidogrel rapidly inhibits platelet aggregation. A loading dose of clopidogrel followed by daily doses of 75 mg has been used to treat patients with acute myocardial ischemia. Suri et al640 administered 600 mg of clopidogrel to 20 patients with a mean interval from stroke of 25 hours. No cases of neurological worsening or intracranial hemorrhage were reported. Another pilot study evaluated the administration of 325 mg of aspirin and 375 mg of clopidogrel to patients within 36 hours of a recent stroke or TIA.641 The combination was found to be safe, and there was a suggestion that neurological deterioration could be prevented. A small Thai study reported the combination of aspirin and dipyridamole also could be administered safely within 48 hours of onset of stroke.642 Overall, these data do not provide solid evidence about the utility of these antiplatelet agents in the management of patients with acute ischemic stroke.
More recently, 2 trials have investigated the early use of antithrombotic drugs in acute stroke. The EARLY trial was an open-label, randomized, multicenter German study of patients with acute ischemic stroke who received 100 mg of aspirin monotherapy or 25 mg of aspirin plus 200 mg of extended-release dipyridamole within 24 hours of stroke or TIA or after 7 days of aspirin monotherapy.643 Of the 543 patients enrolled in both groups, 56% of patients given the combination regimen achieved an mRS of 0 or 1 at 90 days compared with 52% of patients who received aspirin monotherapy. Vascular adverse events, assessed as a composite end point, occurred in 10% and 15% of the early- and late-initiation groups respectively. The Fast Assessment of Stroke and Transient Ischemic Attack to Prevent Early Recurrence (FASTER) pilot trial also recruited patients with ischemic stroke or TIA in a similar study design but only enrolled patients with minor stroke (NIHSS score <4).644 In a factorial design, patients were randomized to clopidogrel or placebo and simvastatin or placebo within 24 hours of their qualifying event. After enrolling 394 patients, the study was stopped prematurely because of the increased use of statins in general. Patients who received clopidogrel had a 90-day stroke risk of 7.1% compared with 10.8% in the placebo arm (adjusted risk ratio, −3.8%; P=0.019). Two patients who received clopidogrel developed intracranial hemorrhage compared with none in the placebo group. These 2 studies suggest that in patients who did receive fibrinolytic therapy, the early initiation of antithrombotic therapy for the secondary prevention of recurrent stroke appears to be as safe as later initiation.
Intravenous Antiplatelet Agents
Inhibitors of the platelet glycoprotein IIb/IIIa receptor are being considered for treatment of acute ischemic stroke because they may increase the rate of recanalization and improve patency of the microcirculation.645,646 A series of studies evaluated one of these agents, abciximab. These included case reports and small clinical series; in some cases, the agent was given as monotherapy and in others as an adjunct, usually with pharmacological fibrinolysis or mechanical thrombectomy.545,635,647–653 Abciximab also was tested in a clinical research program that included a dose-escalation study, a phase II dose-confirmation study, and a phase III clinical trial.654–656 On the basis of the findings of the first 2 studies, the dose and regimen of abciximab used to treat patients with acute coronary lesions were found to have a reasonable safety profile.654,655 In the phase II trial, there was a trend for an improvement in the rate of favorable outcomes among patients treated within 5 hours of stroke.655 Unfortunately, interim analysis of the first 439 patients in the phase III trial did not demonstrate an acceptable risk-benefit ratio for treatment with abciximab, which led to the trial being halted.656 As part of the phase III trial, the investigators also tested the use of abciximab for treatment of patients with stroke present on awakening. The trial found that the risk of bleeding with abciximab in this situation was beyond the desirable safety margins, and the trial halted recruitment of this group in advance of the remainder of the trial.657
Other parenterally administered glycoprotein IIb/IIIa receptor blockers also are being studied as monotherapy or as an adjunct to other recanalization interventions to treat patients with acute ischemic stroke. Most reports involve small series of patients who were treated with either tirofiban or eptifibatide.658–663 Although the use of abciximab to treat acute ischemic stroke caused more hemorrhages, tirofiban did not increase the incidence of cerebral hemorrhagic transformation or parenchymal hemorrhage but may have lowered the mortality rate at 5 months in a phase II trial.664 SaTIS (Safety of Tirofiban in Acute Ischemic Stroke) was a prospective, randomized, placebo-controlled, open-label treatment phase II trial that enrolled 260 patients at 11 centers. In this trial, ischemic stroke patients between 18 and 82 years old with an NIHSS score of 4 to 18 and within 3 to 22 hours of symptom onset were treated with intravenous tirofiban (0.4 µg/kg initial infusion over a 30-minute period, followed by 0.1 µg/kg continuous infusion for 48 hours). Approximately 1% of patients treated developed reversible thrombocytopenia. More patients in the placebo arm were taking aspirin. Of the 3 glycoprotein IIb/IIIa antagonists, tirofiban differs pharmacologically from abciximab and eptifibatide. Perhaps the relatively safer hemorrhagic profile demonstrated in SaTIS is related to tirofiban being a nonpeptide glycoprotein IIb/IIIa antagonist with a biological half-life of 4 to 8 hours and a return of platelet function in 2 hours when stopped.
Recently, the results of the Combined Approach to Lysis Utilizing Eptifibatide and rtPA in Acute Ischemic Stroke (CLEAR) trial were published.532 This randomized, double-blind, dose-escalation study tested the combination of eptifibatide (75 mg/kg bolus and infusion 0.75 mg·kg−1·min−1) and rtPA either 0.3 mg/kg or 0.45 mg/kg IV compared with the conventional dose of intravenous rtPA alone. The study found the combination to be safe, although there was a trend toward better outcomes among those patients who received the conventional dose of intravenous rtPA alone. The investigators are currently conducting a follow-up phase II study, CLEAR-ER.
Most recently, Zinkstok and colleagues665 compared the safety and efficacy of early administration of intravenous aspirin started within 90 minutes after initiation of intravenous rtPA therapy to intravenous rtPA alone in a multicenter, randomized, open-label study. In both groups, oral aspirin therapy was initiated 24 hours after intravenous rtPA. After 642 of a planned 800 patients were enrolled, the trial was terminated prematurely because of an excess of sICH in the aspirin treatment arm. Patients in the combined intravenous aspirin and rtPA group were more than twice as likely to develop sICH as the group given intravenous rtPA alone (4.3% versus 1.6% respectively; P=0.04). There was no significant difference in 90-day outcomes between the combined versus rtPA-alone groups (mRS score 0–2, 57.2% versus 54.0%, respectively).
Conclusions and Recommendations
Currently available data demonstrate a small but statistically significant decline in mortality and unfavorable outcomes with the administration of aspirin within 48 hours after stroke. It appears that the primary effects of aspirin are attributable to a reduction in early recurrent stroke. Data regarding the utility of other antiplatelet agents, including clopidogrel alone or in combination with aspirin, for the treatment of acute ischemic stroke are limited. In addition, data on the safety of antiplatelet agents when given within 24 hours of intravenous fibrinolysis are lacking. The relative indications for the long-term administration of antiplatelet agents to prevent recurrent stroke are included in other guideline and advisory statements.302,666
Research into intravenously administered antiplatelet agents is ongoing. An international trial did not demonstrate an acceptable safety/benefit profile for abciximab when it was administered within 6 hours of acute ischemic stroke. Other agents are being tested in conjunction with mechanical or pharmacological fibrinolysis. Considerably more research is needed to determine whether these agents have a role in the management of patients with acute ischemic stroke.
Oral administration of aspirin (initial dose is 325 mg) within 24 to 48 hours after stroke onset is recommended for treatment of most patients (Class I; Level of Evidence A). (Unchanged from the previous guideline13)
The usefulness of clopidogrel for the treatment of acute ischemic stroke is not well established (Class IIb; Level of Evidence C). Further research testing the usefulness of the emergency administration of clopidogrel in the treatment of patients with acute stroke is required. (Revised from the previous guideline13)
The efficacy of intravenous tirofiban and eptifibatide is not well established, and these agents should be used only in the setting of clinical trials (Class IIb; Level of Evidence C). (New recommendation)
Aspirin is not recommended as a substitute for other acute interventions for treatment of stroke, including intravenous rtPA (Class III; Level of Evidence B). (Unchanged from the previous guideline13)
The administration of other intravenous antiplatelet agents that inhibit the glycoprotein IIb/IIIa receptor is not recommended (Class III; Level of Evidence B). (Revised from the previous guideline13) Further research testing the usefulness of emergency administration of these medications as a treatment option in patients with acute ischemic stroke is required.
The administration of aspirin (or other antiplatelet agents) as an adjunctive therapy within 24 hours of intravenous fibrinolysis is not recommended (Class III; Level of Evidence C). (Revised from the previous guideline13)
Volume Expansion, Vasodilators, and Induced Hypertension
Ischemic stroke results from occlusion of an artery with subsequent reduction in regional cerebral blood flow, demarcated into 2 distinct regions consisting of regional cerebral blood flow alterations: severe reduction (core) and moderate reduction (penumbra).667,668 The penumbra remains viable for hours because some degree of blood flow is sustained through collateral supply and arteriolar dilation.669,670 For >3 decades, investigators have studied interventions aimed at increasing cerebral perfusion in acute ischemic stroke by either improving flow through partially occluded vessels or improving flow through cerebral collateral circulation. These approaches have targeted acute alterations of blood rheology, expansion of blood volume, and increased global or local blood pressure. To date, no acute clinical trial has demonstrated unequivocal efficacy, but several ongoing trials may provide a new, widely applicable therapy for patients with ischemic stroke.
Hypervolemia and Hemodilution for Treatment of Acute Ischemic Stroke
Increased viscosity has been observed in the acute period of ischemic stroke because of volume depletion, leukocyte activation, red cell aggregation, elevated fibrinogen levels, and reduced red cell deformability.671–675 A higher hematocrit is associated with reduced reperfusion, greater infarct size, and higher mortality among patients after ischemic stroke.671,674 Hemodilution and volume expansion are proposed as treatment options to reduce the viscosity of blood, improve flow through collateral channels and microvascular circulation, and increase oxygen-carrying capacity.676–682
A meta-analysis of 18 trials683 in which hemodilution was initiated within 72 hours of symptom onset was reported. A combination of phlebotomy and plasma volume expanders was used in 8 trials, and volume expansion alone was used in 10 trials. The plasma volume expander was dextran 40 in 12 trials, hydroxyethyl starch in 5 trials, and albumin in 1 trial. Hemodilution did not significantly reduce deaths within the first 4 weeks (OR, 1.1; 95% CI, 0.9–1.4) or within 3 to 6 months (OR, 1.0; 95% CI, 0.8–1.2). The proportion of patients with death, dependency, or institutionalization was similar in both groups (OR, 1.0; 95% CI, 0.8–1.2). There was no increased risk of serious cardiac events among patients with hemodilution.
Vasodilatation in Acute Ischemic Stroke
Techniques to promote vasodilation have been studied in acute stroke for >4 decades. Initially, vasodilatation was studied as a way to treat and prevent TIAs. More recently, vasodilation with methylxanthine derivatives, specifically pentoxifylline, propentofylline, and pentifylline, has been evaluated in the setting of acute ischemic stroke. In addition to the vasodilatation, the methylxanthine drugs may also reduce blood viscosity, increase erythrocyte flexibility, inhibit platelet aggregation, and decrease free radical production. Most methylxanthine-class trials have investigated the promotion of vasodilation in the subacute time frame. In a small randomized trial of 110 Chinese patients with acute cortical and lacunar strokes, Chan and Kay684 initiated vasodilation using pentoxifylline in combination with aspirin within 36 to 48 hours from stroke onset and continued for 5 days. At 1 week, there was no difference in outcomes for patients with lacunar stroke between the treatment arms. They did report a statistically significant reduction in morbidity in patients with cortical strokes.684 Subsequent studies have failed to reproduce this effect, and a Cochrane review of the 4 pentoxifylline trials and the 1 propentofylline study found there was not enough available evidence to reliably assess the effectiveness and safety of methylxanthine drugs in acute ischemic stroke.685
Induced Hypertension for the Management of Acute Ischemic Stroke
Increasing the systemic blood pressure may improve regional cerebral blood flow as a result of augmentation of flow through collaterals and arterioles that do not demonstrate an autoregulatory constrictive response to pathological vasodilation.686–690 The clinical response is varied because of variations in collateral formation and preservation of autoregulatory vasoconstriction, systemic blood pressure response, and presence of a penumbra.
Rordorf et al691 retrospectively reviewed a group of patients admitted with the diagnosis of ischemic stroke, of whom 33 were not given a pressor agent and 30 were treated with phenylephrine within 12 hours of symptom onset. There was no significant difference in morbidity or mortality between the 2 groups of patients. In 10 of 30 patients treated with induced hypertension, a systolic blood pressure threshold (mean 156 mm Hg) was identified below which ischemic deficits worsened and above which deficits improved. The mean number of stenotic/occluded arteries was greater in patients with an identified clinical blood pressure threshold for improvement subsequent to induced hypertension. A second pilot study692 used phenylephrine to raise the systolic blood pressure in patients with acute stroke by 20%, not to exceed 200 mm Hg. Of 13 patients treated, 7 improved by 2 points on the NIHSS. No systemic or neurological complications were observed. Marzan et al693 reported the results of induced hypertension (10%–20% of the initial value) using norepinephrine within a mean period of 13 hours after symptom onset. The dose was gradually reduced after 12 hours of administration and terminated when arterial blood pressure remained stable. Early (within 8 hours of initiation) neurological improvement by ≥2 points on the NIHSS was seen in 9 (27%) of 33 patients. Intracranial hemorrhage occurred in 2 patients. Hillis et al694 randomized consecutive series of patients with large diffusion-perfusion mismatch to induced blood pressure elevation (n=9) or conventional management (n=6). Serial DWI and perfusion-weighted MRI studies were performed before and during the period of induced hypertension. Patients who were treated with induced hypertension showed significant improvement in NIHSS score from day 1 to day 3, cognitive score, and volume of hypoperfused tissue. High correlations were observed between the mean arterial pressure and accuracy on daily cognitive tests. Koenig et al695 reported analysis of 100 patients who underwent perfusion-weighted MRI after acute ischemic stroke, of whom 46 were treated with induced hypertension with various vasopressors. The target mean arterial pressure augmentation of 10% to 20% above baseline was achieved in 35% of the 46 treated patients. Compared with 54 patients who underwent conventional treatment, NIHSS scores were similar during hospitalization and discharge, with no clear difference in rates of adverse events. Shah et al696 reported 3 patients who received induced hypertension, not to exceed 180 mm Hg, after partial recanalization using intra-arterial fibrinolysis and noted favorable outcomes and no complications.
The available evidence suggests that a small subset of patients with ischemic stroke in the very acute period may benefit from modest (10%–20%) pharmacological elevation in systemic blood pressure. No clear criteria are validated for selection of such patients, although patients with large perfusion deficits caused by steno-occlusive disease who are not candidates for fibrinolytic and interventional treatments are the best studied, as well as those patients who demonstrate neurological change that correlates with systemic blood pressure changes. A short period (30–60 minutes) of a vasopressor infusion trial may help identify patients who are potential responders to such treatment.
Albumin for Treatment of Acute Ischemic Stroke
Albumin exerts its purported neuroprotective effect by reducing both endogenous and exogenous oxidative stress, maintaining plasma colloid oncotic pressure, and preserving microvascular integrity in focal cerebral ischemia.697 In experimental models of focal ischemia, albumin reduces ischemic brain swelling, improves regional cerebral blood flow, reduces postischemic thrombosis, improves microvascular flow, and supplies free fatty acids to the postischemic brain.672,698,699 In several observational studies,700,701 low serum albumin at admission correlated with higher rates of death and disability among patients with ischemic stroke. Subsequently, the ALIAS (Albumin in Acute Stroke) Pilot Clinical Trial evaluated 6 doses (0.34–2.05 g/kg)702,703 of 2-hour infusion of 25% human albumin beginning within 16 hours of stroke onset in patients with acute ischemic stroke. Eighty-two subjects received albumin, and 42 of those patients also received intravenous rtPA. The only albumin-related adverse event was mild or moderate pulmonary edema in 13% of the subjects, which confirms reasonable tolerability among patients with acute ischemic stroke without major dose-limiting complications. After adjustment for the intravenous rtPA effect, the probability of good outcome (defined as mRS score 0–1 or NIHSS score 0–1 at 3 months) at the highest 3 albumin tiers was 81% greater than in the lower-dose tiers and was 95% greater than in the comparable NINDS rtPA Stroke Trial historical cohort. The intravenous rtPA–treated subjects who received higher-dose albumin were 3 times more likely to achieve a good outcome than subjects receiving lower-dose albumin. The trial suggested that high-dose albumin treatment may be neuroprotective after ischemic stroke, with a synergistic effect between albumin and intravenous rtPA. A large, randomized, multicenter, placebo-controlled efficacy trial, the phase III ALIAS2 Trial,704 compared 2.0 mg/kg of 25% albumin administered over 2 hours with placebo, with treatment initiated within 5 hours of stroke onset. The primary efficacy end point was either an NIHSS score of 0 to 1, an mRS score of 0 to 1, or both at 3 months.704 An interim safety analysis of the first 436 subjects led to modifications in the study design to enhance safety and minimize development of congestive heart failure.705 An exploratory efficacy analysis of the part 1 study data suggested a trend toward favorable outcomes in patients in the albumin arm.706 In the fall of 2012, the study’s data safety and monitoring board stopped recruitment after an interim analysis, and further results from the study are pending.
Mechanical Flow Augmentation
Mechanical methods to increase cerebral perfusion through Willisian and leptomeningeal collaterals offer the prospect of improving cerebral blood flow without the complications of vasopressor pharmacological agents. Data from animal models and from human research demonstrate that aortic occlusion, which is commonly performed by cross-clamping the descending aorta for vascular control during aortic surgery, results in net flow diversion to the cerebral from the lower-extremity circulatory beds, thereby increasing cerebral blood flow.707–715 This evidence generated the development of a catheter-based device with 2 balloons near its distal tip placed in the infrarenal and suprarenal positions in the descending aorta (NeuroFlo device; CoAxia, Maple Grove, MN). After insertion via the femoral artery, the balloons are inflated sequentially up to ≈70% of the diameter of the aortic lumen over a period of 45 minutes to an hour, followed by removal.716 A clinical feasibility study in acute ischemic stroke enrolled 17 patients up to 12 hours after symptom onset and showed an improvement in neurological symptoms in >50% of patients during treatment and at 24 hours.717 A randomized controlled multicenter trial enrolling patients with ischemic stroke within 14 hours of symptom onset was completed in 2010. Results recently published in Stroke failed to show significant differences in clinical outcome, but no issues of safety were noted.718,719 There was a statistically nonsignificant trend in lowering mortality in the treatment group compared with controls (11.3% versus 6.3%, respectively).
Another method that shows potential for augmenting cerebral blood flow is extracorporeal counterpulsation therapy, which is approved for patients with ischemic heart disease who have refractory angina. This therapy is provided by a device that inflates pneumatic cuffs on the lower extremities in sequential fashion during each cardiac cycle to augment diastolic flow in the coronary arteries and improve systolic unloading in the periphery.720 There is also evidence that it may develop and recruit collateral vessels in ischemic myocardium.721 In the cerebral bed, studies have demonstrated extracorporeal counterpulsation–induced diastolic augmentation of flow in the carotid arteries722 and, more recently, the MCAs.723 In addition, a small pilot trial of subacute extracorporeal counterpulsation in the first 2 months after stroke onset was encouraging.724 On the basis of these findings, a randomized dose-ranging trial is ongoing in patients with acute ischemic stroke who are outside the therapeutic time window for intravenous fibrinolysis or endovascular therapy.
Augmentation of cerebral collateral blood flow is a compelling concept that may hold promise in the treatment of acute ischemic stroke. Although the aforementioned treatments appear to warrant further investigation, there are currently no data to support their use in this population of patients.
In exceptional cases with systemic hypotension producing neurological sequelae, a physician may prescribe vasopressors to improve cerebral blood flow. If drug-induced hypertension is used, close neurological and cardiac monitoring is recommended (Class I; Level of Evidence C). (Revised from the previous guideline13)
The administration of high-dose albumin is not well established as a treatment for most patients with acute ischemic stroke until further definitive evidence regarding efficacy becomes available (Class IIb; Level of Evidence B). (New recommendation)
At present, use of devices to augment cerebral blood flow for the treatment of patients with acute ischemic stroke is not well established (Class IIb; Level of Evidence B). These devices should be used in the setting of clinical trials. (New recommendation)
The usefulness of drug-induced hypertension in patients with acute ischemic stroke is not well established (Class IIb; Level of Evidence B). (Revised from the previous guideline13) Induced hypertension should be performed in the setting of clinical trials.
Hemodilution by volume expansion is not recommended for treatment of patients with acute ischemic stroke (Class III; Level of Evidence A). (Revised from the previous guideline13)
The administration of vasodilatory agents, such as pentoxifylline, is not recommended for treatment of patients with acute ischemic stroke (Class III; Level of Evidence A). (Unchanged from the previous guideline13)
Neuroprotection refers to the concept of applying a therapy that directly affects the brain tissue to salvage or delay the infarction of the still-viable ischemic penumbra, rather than reperfusing the tissue. Because many potential neuroprotective therapies are likely safe and potentially efficacious in hemorrhagic as well as ischemic stroke, the ideal neuroprotective therapy would be initiated as early as possible in the course of therapy, including in the prehospital setting, and be continued while other measures are instituted, such as brain imaging followed by fibrinolytic or endovascular revascularization.
Pharmacological agents that limit the cellular effects of acute ischemia or reperfusion may limit neurological injury after stroke. Potential therapeutic strategies include curbing the effects of excitatory amino acids, such as glutamate, transmembrane fluxes of calcium, intracellular activation of proteases, apoptosis, free radical damage, inflammatory responses, and membrane repair. More than 1000 published reports of various experimental neuroprotective treatments for acute stroke exist, culminating in well over 100 clinical trials.725,726 Most clinical trials testing these therapies have produced disappointing results. In some circumstances, treated patients had worse outcomes than did control subjects, or the rates of adverse events were unacceptably high.727 Most of the early neuroprotection studies initiated therapy past the commonly accepted 4- to 6-hour therapeutic window.697 Although some of these clinical studies were small or poorly designed, others have been sufficiently large and methodologically strong to produce important information.728 Newer agents and innovative clinical trial designs that adhere to the STAIR (Stroke Therapy Academic Industry Roundtable) criteria are needed to demonstrate that neuroprotective strategies could be helpful in treatment of stroke.550
Nimodipine is approved for the prevention of ischemic stroke among people with recent aneurysmal subarachnoid hemorrhage.729 Nimodipine was tested in a large number of primary ischemic stroke clinical trials with generally negative results.413,430,730–732 In some cases, outcomes were worse among patients treated with nimodipine than among control subjects.430,732 Presumably, the higher rates of poor outcomes were secondary to the antihypertensive effects of nimodipine.430 Trials of flunarizine, isradipine, and darodipine were also negative.733–735 Although nicardipine is used to treat elevated blood pressure in the setting of stroke, the agent has had limited testing for neuroprotective treatment of the stroke itself.736,737 A meta-analysis published in 2000 of the calcium channel–blocking agents found no evidence that this class of drug is effective in improving outcomes after ischemic stroke.738
Several N-methyl-d-aspartate antagonists have been tested in clinical trials, with largely negative results and increased rates of serious adverse events.739–752 A 2003 systematic review of the excitatory amino acid modulator trials found no rates of improvements in either death or favorable outcomes with treatment.753 Lubeluzole, which is thought to downregulate the glutamate-activated nitric oxide synthase pathway, was tested in several clinical trials, including one that evaluated the combination of the medication and intravenous rtPA.754 Although a pilot study suggested safety and a reduction in deaths, subsequent larger clinical trials found no effects in reducing deaths or improving outcomes after stroke.502,755,756 A subsequent analysis of the trials concluded that there was no evidence for the effectiveness of lubeluzole.757
Several trials tested the efficacy of clomethiazole, a γ-aminobutyric acid agonist, alone or in combination with intravenous rtPA.758 The medication was also used to treat patients with hemorrhagic stroke.759,760 Larger clinical trials failed to demonstrate the efficacy of clomethiazole in improving outcomes after ischemic stroke.758,761–763 A randomized trial of diazepam, another γ-aminobutyric acid agonist, demonstrated no improvement in outcome at 3 months.764 Although a dose-escalation study of naloxone found the medication to be safe, no signal of efficacy was noted.765 Similarly, no benefit was noted in trials of the opioid antagonist nalmefene.472,766
Free radicals produced during cerebral ischemia are well-known mediators of neuronal injury. In initial studies, NXY-059, a free radical–trapping agent, demonstrated tolerability.767 An initial pivotal trial showed potential in improving disability at 90 days, as measured by the mRS, and in reducing rates of intracranial bleeding768; however, a confirmatory pivotal trial in >3000 patients found no benefit on functional status at 90 days or on rates of intracranial hemorrhage.769 A trial of tirilazad, a free radical scavenger agent that inhibits lipid peroxidation, was halted prematurely when an interim analysis failed to detect efficacy.770,771 A review of all trials testing tirilazad, including in the treatment of subarachnoid hemorrhage, concluded that it did not improve outcomes.772 A dose-escalation study of ebselen, an antioxidant, suggested that it might be safe and effective in improving outcomes after stroke.773 A phase III trial completed enrollment in 2002, but no results were reported.774 A small clinical trial found that edaravone, a free radical scavenger and antioxidant, might improve outcomes.775 To date, none of these agents have sufficient data to support their use.
Trials of neuroprotective agents continue. A pilot study testing the combination of caffeine and alcohol when started within 6 hours of stroke found the intervention to be relatively safe.776 Further evaluation of this intervention in combination with intravenous rtPA and with intravenous rtPA plus hypothermia is under way. Magnesium, an excitatory amino acid blocker, calcium channel blocker, and cerebral vasodilator, has been tested in a series of clinical studies. Although preliminary studies showed that magnesium was well tolerated and might improve outcomes, a subsequent larger clinical trial was negative.777–780 One criticism of these early trials was that the agent was given up to 12 hours after onset of stroke. Subsequently, a study tested the safety and feasibility of very early magnesium sulfate administration by paramedics in the field to suspected stroke patients after informed consent was obtained by telephone. Of 20 patients enrolled (80% of whom had ischemic strokes), 70% received magnesium infusion within 2 hours of symptom onset.781 A larger, phase III prehospital magnesium trial is currently under way.
Citicoline, a phospholipid precursor that appears to stabilize membranes, has been tested in several clinical studies.782–784 The trials did not demonstrate treatment efficacy; however, a subsequent study-level meta-analysis suggested a net benefit of treatment in reducing disability.785 A patient-level pooled analysis reported that patients with moderate to severe stroke might be helped if the medication were started within 24 hours of onset of symptoms.786 The International Citicoline Trial on Acute Stroke (ICTUS), a large, European, multicenter randomized trial of citicoline, enrolled 2298 patients with moderate to severe ischemic strokes within 24 hours from symptom onset.787 The trial was stopped prematurely in 2011 because of futility; no difference was found in the 90-day global outcome end point (OR, 1.03; 95% CI, 0.86–1.25; P=0.364).788 Several trials of GM1-ganglioside, which also may stabilize membranes, have not demonstrated improved outcomes with treatment,789–792 and a systematic review of this agent did not demonstrate any benefit from treatment.793
In addition to their low-density lipoprotein cholesterol–lowering effects, statins, or HMG-CoA reductase inhibitors, exert acute neuroprotective properties, including beneficial effects on endothelial function, cerebral blood flow, and inflammation. Formal dose-escalation trials are under way to evaluate statins as acute neuroprotective agents.794 In a small, 89-patient randomized trial, patients already taking chronic statins at the time of ischemic stroke were randomized within 24 hours of onset to statin withdrawal for 3 days or to continued statin therapy. Among enrolled patients, median time from onset to inclusion was 6 hours. Brief withdrawal of statins during the acute period was associated with increased odds of death or dependency at 3 months.795 Further study on the utility of early statin administration is needed.
Hematopoietic growth factors, in addition to regulating bone marrow, exert multiple potentially neuroprotective effects in the human brain. In a small pilot trial, erythropoietin was associated with a nonsignificant reduction in combined death and dependency796; however, preliminary data from a pivotal trial suggested that treatment with erythropoietin increased mortality.797 Another phase I trial of erythropoietin in acute stroke is under way. Granulocyte colony-stimulating factor has been associated with a nonsignificant reduction in combined death and dependency in 2 small trials.798
Medications that reduce the inflammatory response to ischemia have also been evaluated. A randomized trial of enlimomab (an intercellular adhesion molecule-1 antagonist) found that the rates of poor outcomes, including death, were increased among patients who received the agent.799 Another trial tested a neutrophil inhibitory factor; although the medication was safe, it did not improve outcomes.800 A small study of cerebrolysin, with potential neurotrophic and neuroprotective actions, found that it was safe and might improve outcomes.801 Preliminary studies of trafermin (basic fibroblast growth factor) showed that it was well tolerated but that there was a higher death rate among treated patients.728,802 Other potentially neuroprotective therapies that are being tested include interferon-β, adenosine A1 receptor agonists, and nitric oxide synthase inhibitors.
Considerable experimental and clinical research is required before a pharmaceutical agent with identified neuroprotective effects can be recommended for treatment of patients with acute ischemic stroke. Several steps to improve preclinical and clinical research in neuroprotective agents, such as the STAIR guidelines, have been recommended.803,804 It is hoped that ongoing studies of neuroprotective agents, potentially tested alone or in combination with measures to restore perfusion, will demonstrate safety and efficacy.
Hypothermia has been shown to be neuroprotective in experimental and focal hypoxic brain injury models. Hypothermia may delay depletion of energy reserves, lessen intracellular acidosis, slow influx of calcium into ischemic cells, suppress production of oxygen free radicals, alter apoptotic signals, inhibit inflammation and cytokine production, and lessen the impact of excitatory amino acids.805,806 Deep hypothermia is often administered to protect the brain in major operative procedures. Mild to moderate hypothermia is associated with improved neurological outcomes among patients with cardiac arrest, which led to hypothermia becoming the first neuroprotective strategy to be recommended by the AHA in comatose patients after cardiac arrest.807–810 Conversely, a multicenter clinical trial found that mild hypothermia administered during surgery for treatment of a ruptured intracranial aneurysm did not improve outcomes after subarachnoid hemorrhage.811
Several small clinical studies have evaluated the feasibility of inducing modest hypothermia for treatment of patients with acute ischemic stroke.812–818 Two small studies evaluated the utility of hypothermia in treating patients with malignant cerebral infarctions; results were mixed.819,820 Potential side effects of therapeutic hypothermia include hypotension, cardiac arrhythmias, and pneumonia.821 Den Hertog et al,822 in a 2009 systematic review, found no indication of clinical benefit or harm from the use of hypothermia in stroke. A clinically significant effect could not be ruled out, however, and it was advised that large clinical trials were needed to assess the effect of hypothermia.822
Most pilot clinical trials to date have been designed to establish the safety and feasibility of various cooling techniques. These have typically used cohort or case-control groups for comparison of clinical efficacy. To date, no trial has produced Class I evidence, and none has had sufficient sample size to provide robust results. In studies investigating mild to moderate hypothermia induced by use of cooling blankets, the rate of cooling has been relatively slow, and shivering becomes an issue in nonparalyzed, non–mechanically ventilated patients. Moderate hypothermia, especially via endovascular techniques, can reach target temperatures more quickly, but this degree of hypothermia (32°C–33°C) appears to be associated with increased complications, including hypotension, cardiac arrhythmias, pneumonia, and thrombocytopenia. Patients with severe hemispheric strokes, especially with edema and mass effect, appear to be vulnerable to rebound increases in ICP when the rate of rewarming is relatively rapid. Mild or modest hypothermia (34°C–35°C) appears to produce fewer significant clinical complications.
Numerous questions remain unanswered related to the clinical use of hypothermia in acute focal cerebral ischemia. These include the therapeutic window for initiation of hypothermia, the speed of hypothermia induction, the level and duration of hypothermia, the rate of rewarming, and the most effective form of hypothermia delivery with the fewest complications. Additional questions to be addressed include proper or optimal patient selection, concomitant interventions such as fibrinolytics and hemicraniectomy, and whether hypothermia should be regional (cooling helmets or regional hypothermic saline infusions) or systemic (cooling blankets or endovascular catheters).
Lastly, many authors are promoting the investigation of hypothermia in conjunction with other potentially neuroprotective strategies. Synergistic effects of hypothermia with intravenous magnesium, caffeine, and alcohol have been proposed for study.823,824 Ongoing feasibility and larger clinical trials of induced hypothermia, either alone or in combination with other therapies, will likely increase our understanding of the role of hypothermia in acute cerebral ischemia. Until then, there remains insufficient clinical evidence to establish a class of recommendation for induced hypothermia in acute stroke.
Hyperbaric oxygen therapy (HBOT) is delivered in a specialized chamber pressurized to multiples of the ambient atmosphere (atmospheres absolute, or ATA; typically 1.5 to 3.0) and filled with oxygen to percentages up to 100%. This results in increasing the solubility of oxygen in plasma to a level adequate to support tissues with minimal extraction of the oxygen bound to hemoglobin.825 Systemic harmful effects are generally limited to transient myopia, barotrauma of the middle ear or sinuses, and claustrophobia, but occasionally, HBOT may induce seizures.825 HBOT may be used to treat patients with ischemic neurological symptoms secondary to air embolism or decompression sickness.826,827 Although HBOT has generally conferred beneficial effects in preclinical acute cerebral ischemia studies,828–833 clinical trials of HBOT in patients with acute stroke have been inconclusive or have shown that the intervention does not improve outcomes.834–837 A meta-analysis found no evidence that HBOT improves clinical outcomes for acute stroke.838 Delay from stroke onset to initiation of HBOT was an issue in these trials but is an intrinsic problem with HBOT, given the need for care delivery in a specialized chamber. At present, data do not support the routine use of hyperbaric oxygen in the treatment of patients with acute ischemic stroke.
Near-Infrared Laser Therapy
Application of a low-energy laser to the shaved skull to deliver energy in the near-infrared spectrum at a wavelength of 808 nm has been studied as a potential therapy for acute ischemic stroke.839 The postulated mechanism of action is photobiostimulation, with near-infrared radiation absorbed by mitochondrial chromophores, which accelerates enzymatic activity, increases adenosine triphosphate production, and promotes tissue preservation in the ischemic penumbra and enhanced neurorecovery.840–842
Evidence of benefit in animal models843–846 led to a safety and preliminary efficacy trial in 120 patients with acute ischemic stroke, which demonstrated statistically better outcomes in the treated patients as measured by the NIHSS, mRS, Barthel index, and Glasgow Outcome Scale.847 A confirmatory trial enrolling 660 patients reported a positive trend but not a definitive benefit, and an additional pivotal trial using refined selection criteria is planned.848 Thus far, however, the efficacy of near-infrared laser therapy has not been proven in acute ischemic stroke.
Among patients already taking statins at the time of onset of ischemic stroke, continuation of statin therapy during the acute period is reasonable (Class IIa; Level of Evidence B). (New recommendation)
The utility of induced hypothermia for the treatment of patients with ischemic stroke is not well established, and further trials are recommended (Class IIb; Level of Evidence B). (Revised from the previous guideline13)
At present, transcranial near-infrared laser therapy is not well established for the treatment of acute ischemic stroke (Class IIb; Level of Evidence B), and further trials are recommended. (New recommendation)
At present, no pharmacological agents with putative neuroprotective actions have demonstrated efficacy in improving outcomes after ischemic stroke, and therefore, other neuroprotective agents are not recommended (Class III; Level of Evidence A). (Revised from the previous guideline13)
Data on the utility of hyperbaric oxygen are inconclusive, and some data imply that the intervention may be harmful. Thus, with the exception of stroke secondary to air embolization, this intervention is not recommended for treatment of patients with acute ischemic stroke (Class III; Level of Evidence B). (Unchanged from the previous guideline13)
Enthusiasm has grown over the past several years for early and sometimes immediate revascularization (emergent, typically within first 24 hours) with CEA in patients presenting with acute stroke or with stroke in evolution. Justification for this strategy is based on the reported risk of recurrent stroke for patients undergoing medical therapy while awaiting revascularization.849,850
In addition, there are theoretical benefits bestowed by (1) removal of the source of thromboembolic debris (thereby reducing chance of recurrent events, particularly in the case of “soft” or “ulcerated” plaque) and (2) restoring normal perfusion pressure to the ischemic penumbra in the brain. Data suggest that delaying CEA may reduce the potential benefit of revascularization by exposing certain patients to greater risk of recurrent stroke (up to 9.5% in the North American Symptomatic Carotid Endarterectomy Trial).850a Early CEA is believed to reduce that risk. Tempering the enthusiasm for early intervention are concerns regarding transformation of ischemic infarction to hemorrhagic infarction, as well as the potential to increase edema or produce hyperperfusion syndrome from sudden restoration of normal perfusion pressure to the brain. Sbarigia et al851 enrolled 96 patients in a single-arm multicenter trial to evaluate the safety and efficacy of early CEA. Patients with very large ischemic strokes (NIHSS score >22) or with more than two thirds of the MCA territory involved with infarction were excluded. Mean time between onset of stroke and CEA was 1.5 days (±2 days). Overall 30-day morbidity/mortality was 7.3% (7/96). Most patients (85/96) demonstrated significant improvement; only 3% developed greater deficits, and no patients in this carefully selected cohort had hemorrhagic transformation or new cerebral infarction on CT. In another multicenter trial, Ballotta et al852 performed early or urgent CEA (eg, within 2 weeks of acute stroke presentation; median time 8 days) on 102 patients with an mRS score <2. None of the subjects experienced new strokes, hemorrhagic conversions, or cerebral edema. Notably, case selection was limited to those with minor nondisabling stroke, who were neurologically stable, and with limited territorial infarct on CT or MRI. Case series in which more ill or neurologically unstable patients underwent early CEA demonstrated less favorable results. Huber et al853 and Welsh et al854 described combined stroke and death rates of 16% and 21%, respectively; their patients were more neurologically unstable, and some had complete carotid occlusion. Paty et al855 showed that as infarct size increased by 1 cm in diameter, risk of permanent neurological impairment after CEA increased by a factor of 1.7. Thus, it would appear that early CEA may be appropriate for those with small, nondisabling stroke, with the goal of reducing ongoing thromboembolism or flow-limiting ischemia.
A systematic review by Rerkasem and Rothwell856 of outcomes from a large number of publications specifically examined the influence of timing between onset of symptoms of TIA/stroke and subsequent CEA. These authors point out the paucity of data regarding optimal timing of CEA in general and specifically regarding outcomes for CEA for stroke-in-evolution or crescendo TIA. Existing studies have highly variable elements and definitions for these entities, and there is a lack of standardization across studies. Rerkasem and Rothwell’s pooled analysis of results from 47 relevant studies published through 2008 demonstrated relatively high combined stroke and death rates for urgent CEA, 20.2% and 11.4%, in settings of stroke-in-evolution and crescendo TIA, respectively. There was no improvement in outcomes over time, because event rates from studies conducted before and after 2000 were not different. The incidence of stroke and death was significantly higher in patients who required emergent surgery for stroke-in-evolution or crescendo TIA than in patients with nonemergency CEA (OR, 4.6). All but 2 small studies in this analysis excluded patients who had major stroke; most patients had nondisabling stroke or variable deficits (crescendo TIA) as of the time of surgery. Emergent and urgent (days) surgery after large disabling stroke, regardless of carotid status, remains high risk.
Rerkasem and Rothwell856 conclude the following: (1) Risks of emergency CEA are high in patients with unstable neurological status; (2) this risk must be balanced against the risk of neurological deterioration on medical therapy; (3) current evidence does not support emergent CEA for such patients; (4) improvements in intensive medical therapy may allow for stabilization of such patients; and (5) prospective randomized controlled trials of emergent or urgent versus delayed revascularization in patients with unstable neurological status (acute evolving stroke or crescendo TIA) are warranted. In contrast to patients with ongoing instability, data indicate that those patients who are neurologically stable after presenting with a nondisabling stroke or TIA may undergo surgery early on without any incremental risk compared with delayed surgery. Because the incidence of recurrent stroke or TIA is highest early after initial presentation, this subset of patients likely benefits from early revascularization. Data from large randomized trials show that the absolute benefit of CEA is highest during the initial 2 weeks after the event, provided the patient is not demonstrating instability. Rerkasem and Rothwell856 highlight the need for additional carefully designed studies to compare alternative treatment algorithms for patients with acute neurological symptoms.
Most data available regarding the effectiveness of surgical treatment of patients with ischemic stroke or TIA do not pertain to CEA immediately after presentation but rather hours, days, or weeks after the initial event. Few data are available regarding emergency surgical intervention to treat or reverse the initial acute stroke. The most accepted and most common indication for immediate operation for acute stroke is in the setting of a new deficit that occurs immediately after CEA. Surgery in such instances is performed to correct a technical issue that resulted in attenuation of flow or acute thrombosis. Emergency CEA generally is not performed in other settings of acute ischemic stroke, especially when the deficit is large, because of the high risk of adverse events associated with acute restoration of flow to damaged tissue. The exception to this might be when either clinical parameters or DWI suggests that the actual infarcted area is small and the penumbra is large, which indicates that reperfusion of a severe carotid narrowing might enable recovery of tissue in the ischemic zone.
Emergent CEA is sometimes advocated for patients with intraluminal mobile or sessile thrombus associated with an atherosclerotic plaque at the carotid bifurcation. The indications for this are controversial. The morbidity associated with surgery appears to be high among patients who already have intraluminal thrombus demonstrated by cerebral angiography.857–860 Although some groups report low rates of complications and good neurological outcomes with immediate surgery,857–859 others have reported better results when the patients are treated initially with anticoagulants followed by delayed operation.860
Other Surgical Procedures (Extracranial-Intracranial Bypass)
Extracranial-intracranial bypass for the treatment of ischemic stroke has not been shown to be of benefit. Rare reports of improvement with early bypass surgery exist,861,862 as do reports of no improvement and hemorrhagic complications.863 Reports of the early use of surgical embolectomy exist,864,865 but endovascular approaches appear to provide a better alternative in most situations.866,867
Conclusions and Recommendations
Emergent CEA and other operations for treatment of patients with acute ischemic stroke may have serious risks, and the indications must be considered carefully for each individual patient. Furthermore, optimal timing for revascularization after presentation with acute stroke or TIA remains to be defined and likely will vary depending on several factors, including size of infarct, presence and size of residual penumbra, stability of neurological status, and general medical condition of the patient. Additional randomized clinical trials should be designed and undertaken to examine the safety and efficacy of CEA in various subsets of patients with acute stroke, to establish the optimal timing for revascularization, and to define its role in the emergency management of stroke.
The usefulness of emergent or urgent CEA when clinical indicators or brain imaging suggests a small infarct core with large territory at risk (eg, penumbra), compromised by inadequate flow from a critical carotid stenosis or occlusion, or in the case of acute neurological deficit after CEA, in which acute thrombosis of the surgical site is suspected, is not well established (Class IIb; Level of Evidence B). (New recommendation)
In patients with unstable neurological status (either stroke-in-evolution or crescendo TIA), the efficacy of emergent or urgent CEA is not well established (Class IIb; Level of Evidence B). (New recommendation)
Admission to the Hospital and General Acute Treatment (After Hospitalization)
Key to safe and effective stroke care, especially after intravenous or intra-arterial recanalization, is rapid hospital admission or interhospital transfer to a stroke unit or neurocritical care unit. Approximately 25% of patients may have neurological worsening during the first 24 to 48 hours after stroke, and it is difficult to predict which patients will deteriorate.868–872 In addition to the potential progression of the initial stroke, the need to prevent neurological or medical complications also means that patients with acute stroke should be admitted to the hospital in almost all circumstances.873–877 The goals of treatment after admission to the hospital are to (1) observe for changes in the patient’s condition that might prompt initiation of medical or surgical interventions, (2) provide observation and treatment to reduce the likelihood of bleeding complications after the use of intravenous rtPA, (3) facilitate medical or surgical measures aimed at improving outcome after stroke, (4) begin measures to prevent subacute complications, (5) initiate long-term therapies to prevent recurrent stroke, and (6) start efforts to restore neurological function through rehabilitation and good supportive care. The importance of dedicated stroke nursing care in the management of stroke patients cannot be overstated. The 2009 scientific statement from the AHA by Summers et al, entitled “Comprehensive Overview of Nursing and Interdisciplinary Care of the Acute Ischemic Stroke Patient,” is an excellent resource detailing such care.878
Specialized Stroke Care Units
Numerous studies, performed mainly in Europe and Canada, demonstrate the utility of comprehensive stroke units in lessening the rates of mortality and morbidity after stroke.879–892 The positive effects persist for years. The benefits from treatment in a stroke unit are comparable to the effects achieved with intravenous administration of rtPA.893 European stroke units usually do not include intensive care unit–level treatment, including ventilatory assistance. Regular communications and coordinated care are also key aspects of the stroke unit. Standardized stroke orders or integrated stroke pathways improve adherence to best practices for treatment of patients with stroke.894–898 An observational study of New York State stroke data found transport and admission to a PSC compared with nondesignated hospitals led to lower overall 30-day mortality rates (10.1% versus 12.5%) and increased use of fibrinolytic therapy (4.8% versus 1.7%).48 Additional studies have shown that participation in the GWTG-Stroke program has produced improved care processes and sustained increased adherence to stroke performance measures.87,88
Studies demonstrating the benefit of CSCs lag those of PSC effectiveness. An observational study of clinical registries and a linked administrative database in 333 hospitals in Finland demonstrated improved mortality and clinical outcomes when patients were cared for in stroke centers compared with general hospitals.41 The number needed to treat for the prevention of 1 death or institutional care at 1 year was 29 for CSCs and 40 for PSCs compared with nonstroke centers. Prior epidemiological work demonstrated that patients admitted on the weekend had higher mortality. A prospective registry study suggested that CSCs with 24/7 specialized care may ameliorate this occurrence, but additional prospective studies must be performed.899 Given the challenges of building effective stroke systems, continued research is required to identify the best means for triaging patients and integrating nonstroke centers with PSCs and CSCs.
General Stroke Care
Most of the individual components of general medical management after admission to the hospital have not been tested in clinical studies.873,874,876,900,901 Thus, recommendations are based on customary care and the findings from multiple randomized trials that efficient delivery of the combination of these treatments in a stroke unit yields better outcomes than does less organized delivery of these therapies in general medical wards. Medical and nursing management focuses on prevention of subacute complications. Sixty-three percent of patients have ≥1 complication after acute stroke even when cared for in specialized units. The most common complications during the first week in a Norwegian stroke unit were pain, fever, progressing stroke, and UTI. There were low incidences of immobility complications such as DVT and PE in the specialized unit.902 The patient’s neurological status and vital signs are assessed frequently during the first 24 hours after admission. Stroke severity is associated with the development of complications, which most commonly occur in the first 4 days.902 Most patients are first treated with bed rest, but mobilization usually begins as soon as the patient’s condition is considered stable. A Very Early Rehabilitation Trial for Stroke (AVERT) is a large randomized controlled trial that is mobilizing stroke patients within the first 24 hours.903 In the pilot trial, the intervention appeared safe and feasible. Some patients may have neurological worsening on movement to an upright posture. Thus, close observation should be included during the transition to sitting or standing. Early mobilization is favored because it lessens the likelihood of complications such as pneumonia, DVT, PE, and pressure sores.901 In addition, prolonged immobility may lead to contractures, orthopedic complications, or pressure palsies.876,904,905 Frequent turning, the use of alternating pressure mattresses, and close surveillance of the skin help to prevent pressure sores. Measures to avoid falls are also important considerations.906
Nutrition and Hydration
Sustaining nutrition is important because dehydration or malnutrition may slow recovery.907,908 Dehydration is a potential cause of DVT after stroke. Impairments of swallowing are associated with a high risk of pneumonia.909 Some patients cannot receive food or fluids orally because of impairments in swallowing or mental status. Patients with infarctions of the brain stem, multiple strokes, major hemispheric lesions, or depressed consciousness are at greatest risk for aspiration. Swallowing impairments are associated with an increased risk of death.910 An abnormal gag reflex, impaired voluntary cough, dysphonia, incomplete oral-labial closure, a high NIHSS score, or cranial nerve palsies should alert the care team to the risk of dysphagia.9-1-1–913 A preserved gag reflex may not indicate safety with swallowing.914 The patient may be placed on a strict nothing-by-mouth order until an assessment of the ability to swallow is completed. Studies have shown that other healthcare providers can safely perform the initial screening before the speech language pathologist assessment.903,915,916 In a prospective 15-hospital study, use of a formal dysphagia screening protocol, which incorporated an evidence-based screening tool, was associated with improved compliance with dysphagia screenings and a significantly reduced risk of pneumonia.917 The Toronto Bedside Swallowing Screening test, an evidence-based tool for swallow assessment, has been evaluated successfully for interrater reliability and predictive validity.918 A water swallow test performed at the bedside is a useful screening tool. A wet voice after swallowing is a predictor of a high risk for aspiration. Clinical signs may not identify patients at risk for aspiration, and further testing, including a video fluoroscopic evaluation of swallow or a fiber optic endoscopic evaluation of swallow, may be performed if indicated.919–921
Most patients are treated initially with intravenous fluids. Intravenous hyperalimentation is rarely necessary. When necessary, a nasogastric (NG) or nasoduodenal tube may be inserted to provide feedings and to facilitate administration of medications.922 Placement of a percutaneous endoscopic gastrostomy (PEG) tube is performed to treat patients who will need prolonged tube feedings.923 Although this device usually requires less care, complications, including involuntary removal of the tube or peritonitis, may occur.924 The risk of aspiration pneumonia is not eliminated by the use of an NG or PEG tube.
The Feed or Ordinary Diet (FOOD) trials examined (1) the effect of administration of nutritional supplements in outcomes of patients with stroke who could swallow, (2) the effect of initiation of NG feeding started within 7 days of stroke compared with later intervals on outcomes, and (3) the effect of PEG feedings on outcomes compared with NG feedings during the first 2 to 3 weeks after onset.925–928 The results showed that supplemental nutrition was not necessary, that early NG tube feeding may substantially decrease the risk of death, and that early feeding via an NG tube resulted in better functional outcomes than feeding by PEG,925,926 although many long-term care facilities will not accept patients with an NG tube as the means for providing nutrition.
Bowel management to avoid constipation and fecal impaction or diarrhea is also a component of ancillary care.929 Constipation occurs in 30% to 60% of patients 4 weeks after stroke, and in patients with moderate stroke severity, constipation was associated with poor outcomes at 12 weeks.930 Some feedings administered via a PEG or NG tube may cause osmotic gradients that lead to diarrhea.
Pneumonia, which is most likely to occur in seriously affected, immobile patients and those who are unable to cough, is an important cause of death after stroke.876,909,931–933 Aslanyan et al931 found that the development of pneumonia was associated with an increased risk of death (hazard ratio, 2.2; 95% CI, 1.5–3.3) or unfavorable outcome (OR, 3.8; 95% CI, 2.2–6.7). Stroke-associated pneumonia increases length of stay, mortality, and hospital costs.934 Immobility and atelectasis can lead to development of pneumonia. Early mobility and good pulmonary care can help prevent pneumonia.934 Preventive measures in intubated patients include ventilation in a semirecumbent position, positioning of the airway, suctioning, early mobility, and shortened use of intubation if feasible.935 Measures to treat nausea and vomiting may also lower the risk of aspiration pneumonia. Exercise and encouragement to take deep breaths may help to lessen the development of atelectasis. The appearance of fever after stroke should prompt a search for pneumonia, and appropriate antibiotic therapy should be administered promptly. In one study, prophylactic administration of levofloxacin was not successful in lessening the risk of pneumonia or other infections in the first days after stroke.936
UTIs are common, occurring in 15% to 60% of stroke patients; they independently predict worse outcomes and can lead to bacteremia or sepsis as a potential complication.874,931,937–939 Urinalysis for evidence of infection should be performed whenever a patient develops a fever after stroke. Some patients, especially those with major impairments, are at high risk for urinary incontinence.940 Indwelling catheters should be avoided if possible but may be required in the acute phase of stroke. The catheter should be removed as soon as the patient is medically and neurologically stable. Intermittent catheterization may lessen the risk of infection. External catheters, incontinence pants, and intermittent catheterization are alternatives to an indwelling catheter. The patient should be assessed for UTI if there is a change in level of consciousness and no other cause of neurological deterioration is identified. A urinalysis and urine culture should be obtained if UTI is suspected.13,931,932,940 Acidification of the urine may lessen the risk of infection, and anticholinergic agents may help in recovery of bladder function. Although prophylactic administration of antibiotics usually is not done, appropriate antibiotics should be prescribed for patients with evidence of UTI.
DVT and PE
PE accounts for 10% of deaths after stroke, and the complication may be detected in 1% of patients who have had a stroke.941 Indredavik and colleagues902 found PE in <2.5% of patients during the first week in a specialized stroke unit. DVT and PE were more likely to occur in the first 3 months after stroke, with an incidence of 2.5% and 1.2%, respectively.902 Pulmonary emboli generally arise from venous thrombi that develop in a paralyzed lower extremity or pelvis. Besides being associated with a life-threatening pulmonary event, symptomatic DVT also slows recovery and rehabilitation after stroke. The risk of DVT is highest amongst immobilized and older patients with severe stroke.942–946
The options for lowering the risk of DVT include early mobilization, administration of antithrombotic agents, and the use of external compression devices. Anticoagulants are given to prevent DVT and PE among seriously ill patients. Much of the support for the use of anticoagulants comes from clinical studies testing these agents in the treatment of bedridden patients other than those with stroke.947,948 A meta-analysis demonstrated that these medications were effective among patients with stroke.949 Several clinical trials have demonstrated the utility of heparin and LMWH.947,950 The results of the PREVAIL Trial showed that a 40-mg injection of enoxaparin once daily was more effective than 5000 IU of UFH twice a day for prevention of DVT in ischemic stroke patients.629 The risk of serious bleeding complications was relatively low.951 Long-term treatment usually involves the use of oral anticoagulants such as warfarin. Ridker et al952 found that low-intensity warfarin therapy was effective in preventing recurrent venous thromboembolism. Aspirin also may be effective for patients who have contraindications to anticoagulants, although direct comparisons with anticoagulants are limited.606,953,954 Experience evaluating the use of external compression of the veins in the lower extremities, such as stockings or alternating pressure devices, in stroke patients is limited, and potential for skin damage is a concern.955–957 Patients with PE from thrombi in the lower extremities and a contraindication for antithrombotic treatment may require placement of a device to filter the inferior ven