Text Size:    +   -

Emergency Department Evaluation and Management

5th Edition
June 2015

The Canadian Stroke Best Practice Recommendations for Hyperacute Stroke Care, 5th Edition (2015) is published in the International Journal of Stroke (IJS) and available freely online. To access the specific recommendations for Emergency Department Evaluation and Management and all other sections of the Hyperacute Stroke Care recommendations, please click on this URL which will take you to the recommendations online in the IJS: http://onlinelibrary.wiley.com/doi/10.1111/ijs.12551/full

For the French version of these recommendations, open the appendix at this link :  http://onlinelibrary.wiley.com/store/10.1111/ijs.12551/asset/supinfo/ijs12551-sup-0001-si.zip?v=1&s=cdf3d494242426450aaa522f104ace17857f037a

All other supporting information, including performance measures, implementation resources, evidence summaries and references, remain available through www.strokebestpractices.ca, and not through the IJS.  Please click on the appropriate sections on our website below for this additional content.


Patients who present to hospital with suspected stroke often also have significant physiological abnormalities and comorbidities. These can complicate management of stroke. Signs and symptoms that may explain the cause of the stroke or predict later complications (such as space-occupying infarction, bleeding, or recurrent stroke) and medical conditions such as hypertension or the presence of a coagulopathy, will have an impact on treatment decisions. An efficient and focused assessment is required to understand the needs of each patient.

It is impossible to reliably differentiate infarct from hemorrhage by clinical examination alone. Brain imaging is required to guide management, including the selection of time-sensitive acute stroke treatments. A CT scan or magnetic resonance (MR) imaging is essential to differentiate between ischemic stroke and intracerebral hemorrhage, and stroke mimics, since clinicians may disagree on the clinical diagnosis of stroke (versus not stroke) in about 20 percent of patients.

Initial management of elevated blood pressure in acute stroke patients remains controversial due to the lack of evidence to clearly guide practice. At the same time, this is an area where clinicians often seek guidance from stroke specialists. The recommendations for this area emphasize caution and diligence in monitoring and treating extremely high blood pressure in the first hours after stroke onset.

Diabetes is a major modifiable risk factor for vascular disease that may be first diagnosed at the time of a stroke. Severe hyperglycemia (blood glucose greater than 22 mmol/L) is a relative contraindication to the administration of intravenous tPA. Hyperglycemia at the time acute stroke is associated with increased size of the infarcted area in experimental animals, a greater risk of symptomatic hemorrhage after intravenous tPA treatment, and is associated with poor clinical outcomes in epidemiological studies.

System Implications

  • Local protocols to ensure all stroke patients have rapid access to computed tomography (CT) with CT angiography (CTA) of the extracranial and intracranial vessels completed at the same time as the initial brain imaging.
  • Protocols for ‘code stroke’ activation of the stroke team and diagnostic services prompted by receiving prenotification by paramedics of an incoming suspected stroke patient.
  • Agreements to ensure patients initially managed in rural hospitals without neurovascular imaging capability have timely access to CTA with imaging of the extracranial and intracranial vessels at partnering hospitals.
  • Protocols and standing orders to guide initial blood work and other clinical investigations.
  • Local protocols, especially in rural and remote locations, for rapid access to clinicians experienced in interpretation of diagnostic imaging, including access through telemedicine technology.

Performance Measures

  1. Median time from patient arrival to hospital to first imaging scan.
  2. Median time from patient arrival to hospital to first CTA of extracranial and intracranial vessels.
  3. Proportion of stroke patients who receive initial brain imaging (either CT or CTA) within 30 minutes of hospital arrival for those patients who arrive within acute stroke treatment times.
  4. Proportion of stroke patients who receive a brain CT/CTA within 24 hours of hospital arrival (core).
  5. The proportion of patients with carotid territory events who undergo carotid imaging in the ED.
  6. The proportion of patients who do not have carotid imaging in the ED but who have arrangements made for carotid imaging as an outpatient.
  7. The median time from CBC, INR and thrombin time, Cr/eGFR draw to having results available.
  8. Proportion of patients with blood glucose levels documented during assessment in the emergency department.
  9. Proportion of stroke patients who receive a CT scan in less than 25 minutes from hospital arrival in patients arriving less than 3.5 hours from last known well time, and without contraindications to thrombolysis.
  10. Median time from stroke symptom onset to carotid imaging.

Measurement Notes

  • Data may be obtained from laboratory reports or patient chart.
  • CT and CTA imaging time should be based on time of first slice by the scanner. Specify in your results which type of scan (CT or CTA, separately or combined) was being measured and reported
  • Stratify analysis for patients who arrive within 3.5 hours of stroke symptom onset and those who arrive within 4.5, 6 and 12 hours from stroke symptom onset.
  • Performance measure 1: apply to patients who may be candidates for acute thrombolysis (i.e. who arrive at hospital within 3.5 hours of stroke onset) and for patients who may be eligible for other time-sensitive interventions.
  • Performance measures 1 and 2: Time interval measurements for CT and MRI should be calculated from the time the patient enters the emergency department until the time noted on the actual brain imaging scan.
  • Performance measure 3: For outpatient carotid imaging, a notation should appear in the discharge summary, or in nursing notes, with an indication that the test has actually been requested or requisitioned prior to the patient leaving the hospital.
  • Performance measure 5: Use medical history to determine whether patient was known to have diabetes prior to the stroke event.

Implementation Resources and Knowledge Transfer Tools

Health Care Provider Information

Patient Information

Summary of the Evidence, Evidence Tables and References

Evidence Table 3 Emergency Department Evaluation

Initial Assessment

Patients require immediate evaluation when presenting to the emergency department (ED) with suspected stroke or transient ischemic attack (TIA). For those patients presenting with TIA, their risk for imminent stroke (i.e. within one week) can be evaluated, and investigations/treatment initiated to prevent a future stroke. The accuracy of a variety of clinical decision-making tools to assess stroke risk in patients with TIA, such as the ABDC, ABCD2, ABCD3 and the Canadian TIA score have been evaluated previously. Purroy et al. (2012) evaluated 8 different tools and reported that ABCD3 and ABCD3V were the best predictors of stroke at 7 and 90 days. The corresponding areas under the ROC curve (AUC) were 0.66 (p=0.004) and 0.69 (p<0.001) at day 7 and 0.61 (p=0.015) and 0.63 (p=0.003), at day 90. All other tools, including the California Risk Score, ABCD, ABCD2, ABCDI, ABCD2I, SPI-II and ESRS were unable to predict stroke risk beyond chance alone (p>0.05) at either days 7 or 90. Perry et al. (2014) identified 13 independent predictor of stroke recurrence within 7 days and used them to develop the Canadian TIA Score. The AUC for this tool was 0.77 (95% CI 0.73-0.82). The strongest predictors of stroke were established antiplatelet therapy, initial diastolic blood pressure ≥110 mm Hg, and initial blood glucose ≥15 mmol/L.

Standard assessments for patients with suspected acute stroke include a neurological examination, monitoring of vital signs, blood work, imaging and cardiovascular investigations, dysphagia screens and seizure assessment. Dysphagia screens and seizure management are other important components of initial evaluation of patients in the emergency department. Dysphagia screening is particularly relevant for identifying patients at high risk of aspiration and pneumonia. Lakshminarayan et al. (2010) reported that patients who did not receive a swallowing screen were at higher risk of developing pneumonia compared to patients who received and passed screening (OR=2.2, 95% CI 1.7 to 2.7). Further information regarding the evaluation, assessment and management of dysphagia can be found in Section 5.0: Stroke Rehabilitation. The incidence of early seizure (within 24 hours) following stroke is estimated to be 2% to 6%. Early seizure activity has been shown to be a marker of increased stroke severity (Procaccianti et al. 2012). Cortical involvement and hyperglycemia have also been found to be independent predictors of early seizure activity (Lamy et al. 2003, Procaccianti et al. 2012). Evidence regarding the management of post-stroke seizures is limited. Further information regarding seizure management can be found in Section 4.0: Acute Stroke Management.

Neurovascular Imaging

Immediate access to brain and vascular imaging is required for all patients arriving to hospital with suspected stroke or TIA. A non-contrast CT scan is considered the imaging standard to be used initially to identify acute ischemic stroke and to rule out intracranial hemorrhage. CT scans are quick to perform, easy to tolerate, and are known to be very reliable for the detection of intracerebral hemorrhage. Early detection of hemorrhage is essential since the presence of blood in the brain or subarachnoid space is the main contraindication for the administration of aspirin, anticoagulants and thrombolytic therapy. Early imaging is particularly important for patients who may be potential candidates for thrombolytic therapy, since it has a narrow therapeutic window for administration. Wardlaw et al. (2004) found that a computed tomography (CT) scan for all patients with suspected stroke on admission to hospital was the most cost-effective strategy, despite the increased cost of scans being performed during “off hours”. The higher costs were offset by savings realized through decreased lengths of hospital stay.

CT angiography (CTA) should be performed as part of the initial acute stroke CT imaging protocol. It is fast, simple and helps to identify patients with small core infarcts (ASPECTS 6 or higher) in the anterior circulation, who should be considered for endovascular therapy. Either multiphase or dynamic CTA is recommended over single phase CTA, as the former can be used to assess for both intracranial arterial occlusion and also pial arterial collateral circulation (Menon et al. 2015). Evidence of adequate pial collaterals may predict better response to reperfusion and outcomes in acute ischemic stroke patients (Christoforidis et al. 2005).   CTA is well-tolerated with a very low risk of allergic reaction or renal impairment from contrast administration, and does not pharmacologically interact with t-PA.

CT perfusion (CTP) is another advanced CT imaging modality that can be used to determine infarct core size (based on cerebral blood volume [CBV] maps) and ischemic penumbra (using cerebral blood flow [CBF] or time maps). CTP has been used in recent trials of endovascular therapy to identify patients who were candidates for treatment. In the EXTEND-IA trial, (Campbell et al. 2015), inclusion required a 20% mismatch between core infarct and ischemic penumbra identified using CTP. Due to variability in vendor software, specific CBV volume cut-offs for core infarct size is not standardized. The use of CTP for acute stroke patients should be reserved for centres with well-established CTP protocols and experience in interpreting CTP, or the use of quantitative CTP using RAPID software, and must not substantially delay decisions for acute stroke treatments.

While CT scans are recommended for initial brain imaging following stroke, there are cases where magnetic resonance imaging (MRI) with diffusion-weighted sequences (DWI) may be superior. MRI has been shown to be more has been sensitive in detection of the early changes associated with ischemia, especially in patients with small infarcts. Using the results from 8 studies, Brazzelli et al. (2009) reported that the sensitivity of magnetic resonance imaging (MRI) may be higher than CT scans for the identification of ischemic stroke (99% vs. 39%), although the authors questioned the generalizability of their findings. If an MRI is available and performed in place of CT, enhanced imaging in the form of DWI, GRE and FLAIR is indicated. Brunser et al. (2013) included 842 patients admitted to the emergency department with a suspected ischemic stroke. Diffusion-weighted imaging (DWI) examinations were performed for all patients. For patients with a final diagnosis of stroke, the sensitivity of DWI in detecting ischemic stroke was 90% (95% CI 87.9 to 92.6), and specificity was 97% (95% CI 91.8 to 99.0).

Cardiovascular Investigations

An electrocardiogram (ECG) should be performed immediately to identify arrhythmias for all patients with stroke and TIA presenting to the emergency department. Atrial fibrillation (AF) is commonly diagnosed post-stroke, and is of particular concern due to its role in forming emboli. Sposato et al. (2015) included the results from 11 studies in which cardiac monitoring was initiated in the ED. An estimated 7.7% of patients, without a history of AF, were newly diagnosed. Suissa et al. (2012) included 946 patients with ischemic stroke without history of AF and found that the odds of detection was greatest within the first 24 hours of stroke (OR= 9.82; 95% CI 3.01 to 32.07). Patients who received continuous cardiac monitoring group were more likely to be identified with AF compared with those who received a baseline ECG, 24-hour Holter monitor and additional ECGs when necessary (adj OR= 5.29; 95% CI 2.43 to 11.55). Regardless of the type of monitoring used, the initial ECG will not always detect all cases of AF. In the same study, it was found that ECG monitoring beyond the baseline assessment resulted in the identification of additional cases of AF in 2.3%-14.9% of the population (Suissa et al. 2012). The use of serial ECG assessments over the first 72 hours following stroke can be an effective means of diagnosing AF. For example, Douen et al.(2008) reported there was no significant difference in detection rates between cardiac monitoring groups. AF was identified in 15 new patients using serial ECG and in 9 new patients using a Holter monitor. The majority of these cases were identified within 72 hours (83%).

The use of a transesophageal echocardiography (TEE) is indicated when there is suspected cardiac embolism involvement. For patients with an unknown cause of stroke following baseline diagnostic assessments, and no contraindications to anticoagulation therapy, TEE was found to identify possible sources of cardiac embolism (de Bruijn et al. 2006). In 231 patients with recent stroke (all types) or TIA, TEE was found to perform significantly better than transthoracic echocardiography (TTE) in identifying possible sources of cardiac embolism (55% vs. 39%). Among the 39 patients ≤45 years, a potential cardiac source was identified in 13 patients. Of these, the abnormality was identified by TEE in 10 cases and in 3 cases using TTE. Among 192 patients >45 years, a potential cardiac source of embolism was identified in 59% of patients. TEE confirmed the potential cardiac source in 34 patients, but also detected a potential cardioembolic source in an additional 80 patients.

Acute Blood Pressure Management

There is no evidence to suggest that interventions to increase or decrease blood pressure with vasoactive agents help to improve stroke outcome. In the CATIS trial (He et al. 2014), 4071 patients with acute ischemic stroke were randomized to receive or not receive antihypertensive therapy during hospitalization. Although mean systolic blood pressure was significantly lower among patients in the intervention group, treatment was not associated with significant reduction in the risk of death or major disability at either 14-days (OR= 1.00, 95% CI 0.88 to 1.14) or 3-months (OR= 0.99, 95% CI 0.86 to 1.15) following study entry. Two Cochrane reviews have examined the potential benefits of artificially raising and lowering blood pressure with vasoactive drugs within the first week of stroke. One of the reviews was restricted to the inclusion of RCTs, and included the results from 12 trials (Geeganage & Bath, 2008), while the other included non RCTs as well (Geeganage & Bath, 2010). In both of these reviews, the focus of the majority of the included studies was blood pressure reduction. Treatment was associated with significant early and late reductions in SBP and DBP, but was not associated with significant reduction in the risk of death or a poor outcome within one month, or the end of follow-up. However, the use of vasoactive drugs used to raise blood pressure significantly increased in the odds of death or disability at the end of the trial (OR= 5.41; 95% CI 1.87 to 15.64) (Geeganage & Bath, 2010). Further evidence from a metaregression study (Geeganage & Bath, 2009), which included the results from 37 trials, also suggests that large changes in blood pressure in the early post-stroke period are associated with an increased risk or death and the combined outcome of death/dependency. While the authors also suggested that a decrease in blood pressure between 8mmHg and 14.6mmHg was associated with the lowest odds of poor outcome (death, dependency and intracerrbral hemorrhage), the results were not statistically significant. (Geeganage & Bath, 2009).

For patients treated with thrombolysis, reductions in blood pressure may be indicated. Using the results of 11080 patients included in the SITS-ISTR study who were treated with thrombolysis, Ahmed et al (2009) reported that high systolic BP, 2 to 24 hours after thrombolysis was associated with worse outcome (p>0.001). Blood pressures greater than 170 mmHg were associated with a higher odds of death, dependency and subsequent hemorrhage compared to blood pressures between 141 and 150 mmHg.

Glucose Management

Baseline hyperglycemia has been identified as independent predictor of poor stroke outcome and may be a marker of increased stroke severity. The presence of hyperglycemia may be of particular concern among patients without a history of premorbid diabetes. Using patient data from the ECASS II trial, Yong & Kaste (2008) examined the association between stroke outcomes and four patterns of serum glucose over the initial 24-hour period post stroke. Among 161 patients with pre-morbid diabetes, the odds of poor outcome were not increased significantly for patients with persistent hyperglycemia, or among patients with hyperglycemia at 24 hours, compared with patients with persistent normoglycemia. However, among 587 nondiabeteics, patients with persistent hyperglycemia experienced significantly worse outcomes compared to those with persistent normoglycemia. The odds of a good functional outcome at 30 days, minimal disability at 90 days or neurological improvement over 7 days were significantly reduced compared with patients with persistent normoglycemia, while the odds of 90-day mortality and parenchymal hemorrhage were increased significantly. Since initial hyperglycemia has been associated with poor stroke outcome, several trials have evaluated the potential benefit of tight blood glucose control early following stroke. The largest such study was the GIST-UK trial (Gray et al. 2007) in which 899 patients were randomized to receive variable-dose-insulin glucose potassium insulin (GKI) to maintain blood glucose concentration between 4-7mmol/L or saline (control) as a continuous intravenous infusion for 24 hours. For patients in the control group, if capillary glucose > 17 mmol/L, insulin therapy could be started, at the discretion of the treating physician. Treatment with GKI was not associated with a significant reduction in 90 day mortality (OR= 1.14; 95% CI 0.86 to 1.51; p=0.37) or the avoidance of severe disability (OR= 0.96; 95% CI 0.70 to 1.32). Rescue dextrose was given to 15.7% of GKI-treated patients for asymptomatic prolonged hypoglycaemia. The trial was stopped prematurely due to slow enrolment. More recently, Rosso et al. (2012) randomized 120 patients to receive intravenous administration of insulin (IIT) on a continuous basis or subcutaneous administration (every 4 hours) for 24 hours (SIT). The stop point for treatment was <5.5 mmol/L in the IIT group and 8 mmol/L in the SIT group. Although a significantly higher number of patients in the IIT group achieved and maintained a mean blood glucose level of <7mmol/L, the mean size of infarct growth was significantly higher among patients in the IIT group (27.9 vs. 10.8 cm3, p=0.04), there were significantly more asymptomatic hypoglycemia events among patients in the IIT group (8 vs. 0, p=0.02) and there was no significant difference in the number of patients who experienced a good outcome (45.6% vs. 45.6%) or death (15.6% vs. 10.0%) at 3 months. In a Cochrane review (Bellolio et al. 2014) used the results of 11 RCTs including 1583 adult patients with blood glucose level of > 6.1mmol/L obtained within 24 hours of stroke, Blood-glucose-lowering treatment was not associated with reductions in death or dependency (OR=0.99, 95% CI 0.79-1.2) or final neurological deficit, but treatment did increase the risk of was associated symptomatic and asymptomatic hypoglycemia events.