As a group, neurologic conditions represent a substantial portion of emergency department (ED) visits. Cerebrovascular disease, headache, vertigo and seizures are all common reasons for patients to seek care in the ED. Patients being treated for each of these conditions are amenable to care in an ED observation unit (EDOU) if they require further diagnostic or therapeutic interventions beyond their ED stay. EDOUs are the ideal setting for patients who require advanced imaging such as MRIs, frequent neuro checks or specialist consultation in order to determine if they require admission or can be discharged home.
Key points
- •
ED Observation Units are the ideal setting for continued monitoring, and diagnostic testing for patients presenting to the ED with neurologic complaints.
- •
A transient ischemic attack (TIA) is a sentinel event for stroke. Patients with TIA should be risk stratified and begin risk lowering therapies.
- •
Similarly, the EDOU can be used for further risk stratification of patients with non-disabling cerebrovascular accidents (CVA).
- •
The EDOU is appropriate for patients with other neurologic conditions such as headache, seizures and vertigo. Goals can include neuro checks, imaging or specialist consultation.
Carol Brown is a 58-year-old woman who presents to the emergency department with pronounced weakness for 12 minutes that resolved 1 hour before arrival. She has a history of hypertension and hyperlipidemia, but no prior stroke or transient ischemic attack. Her examination is normal, including a National Institutes of Health Stroke Score of zero. In the emergency department her cardiac monitor shows a sinus rhythm, and her noncontrast head computed tomography scan and blood work is normal. Her ABCD 2 score is 3. In the observation unit, her serial examinations and cardiac monitoring are normal, her echocardiogram and MRI of the brain are normal; however, her magnetic resonance angiogram of the neck vessels shows greater than 70% carotid stenosis. She was admitted to the hospital and underwent successful endarterectomy 5 days later following preoperative clearance. On 1-month follow-up she was asymptomatic and doing well.
Introduction
As a group, neurologic conditions represent a substantial portion of emergency department (ED) visits. In 2011, 4.7% of ED visits were for diseases of the nervous system. Cerebrovascular disease was the fifth leading principal hospital discharge diagnosis for admitted ED patients. Between 2001 and 2011 the percentage of patients receiving neuro imaging in the ED increased 39% (from 66% to 92%). Similarly, headache (HA) is the fourth leading principal reason for an ED visit overall, accounting for 4.3 million visits or 3.2% of all visits.
Transient ischemic attack
Background
In the United States, stroke is the fifth leading cause of death. With roughly 800,000 strokes occurring annually, stroke is also a leading cause of disability, costing the United States roughly $34 billion per year. Eighty-five percent of strokes are ischemic, and between 15% and 30% of ischemic strokes are preceded by a transient ischemic attack (TIA). An estimated 240,000 TIAs occur annually. TIA has been redefined by the American Stroke Association (ASA) as a “(t)ransient episode of neurologic dysfunction caused by focal brain, spinal cord, or retinal ischemia without acute infarction.” Notable updates in this definition include the addition of “infarction” found on brain imaging and the removal of time elements such as 24 hours or 1 hour. Symptoms lasting more than 1 hour have a less than 1 of 6 chance of not having a stroke. Patients with clinical resolution of stroke symptoms that show brain infarction on computed tomography (CT) imaging have a fourfold higher risk of subsequent stroke, and those with a new infarct on MRI have a fivefold higher incidence of subsequent stroke.
Although patients with TIA present with no residual deficits, they are at significant risk of subsequent vascular events. Overall, their approximate 90-day stroke rate is 12%, with half of these strokes occurring within the first 1 to 2 days after a sentinel TIA. In fact, this 90-day stroke rate is more than twice that of patients who have suffered an actual stroke (4%), suggesting that patients with TIA present in a state of greater vascular vulnerability. Yet this is confounded by the fact that these patients appear clinically normal, and most will not have an adverse event during or shortly after their TIA, making TIA an ideal condition for an ED observation unit (EDOU).
Emergency Department Management
The initial ED evaluation and subsequent EDOU management of patients with TIA is focused on the early detection and prevention of stroke, the detection of alternate causes of TIA symptoms, and the prevention of future strokes. This includes a history that includes symptom duration, vascular risk factors, and other potential etiologies. Ideally, the examination should include a structured stroke examination, such as the National Institutes of Health Stroke Score (NIHSS), to improve the clinical detection of small or subtle strokes. Obtaining an electrocardiogram and placing the patient on cardiac monitoring is important for the detection of atrial fibrillation. Blood testing should include a complete blood count, serum glucose, and electrolytes. A prothrombin time is often ordered, as well, in the event that the patient subsequently develops stroke requiring reperfusion therapy. Brain imaging, with either CT or MRI, should be obtained if possible. Although the safety of delaying imaging is unknown, there are several benefits to timely imaging.
The transient ischemia that occurs in patients with TIA may be due to a number of different vascular events ( Table 1 ). Causes of TIA may be broadly classified as intracranial vascular, extracranial vascular, and cardioembolic. When an etiology is not identified, the etiology is classified as cryptogenic. Although this refers to patients with stroke, the testing for causative etiologies for TIA is similar.
Etiology | Causes |
---|---|
Intracranial vascular | Atherosclerotic Branch occlusive disease |
Extracranial vascular | Carotid plaque with arteriogenic emboli Flow-limiting stenosis |
Cardioembolic | Paroxysmal atrial fibrillation Paradoxic embolism Valve disease with clot formation Intramural clot formation associated with dyskinesis |
Cryptogenic | Cardiac etiologies: Paroxysmal atrial fibrillation, subacute bacterial endocarditis (culture negative), papillary fibroelastoma Subtle arterial dissections Central nervous system vasculitis Hypercoagulable state in setting of cancer Metabolic disorders: hyper-homocysteine, Fabry disease Plaque in aortic arch Human immunodeficiency virus, central nervous system infection Drug abuse Genetic etiologies (eg, cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy [CADASIL], mitochondrial encephalomyopathy, lactic acidosis, and stroke-like episodes [MELAS]) |
It is important to consider conditions that commonly mimic TIA. These include epilepsy, postictal paralysis, complex migraines, such as hemiplegic or ocular migraines, cervical disc disease, transient global amnesia, intracranial lesions, such as subdural hematomas and mass lesions, inner ear disease with vertigo, and metabolic derangement of the serum glucose, sodium, or calcium.
A unique challenge with the diagnosis of TIA is the lack of a clear gold standard for this diagnosis. When a neurologist is used as the gold standard, false-positive diagnostic rates range from 25% to 38%. There is even poor agreement between fellowship-trained stroke neurologists.
To improve diagnostic performance, TIA risk stratification scores have been developed. The most common is the ABCD 2 score ( Table 2 ). American Heart Association (AHA)/ASA guidelines recommend admission or discharge based on a total score greater or less than 3. Like other clinical prediction tools, the ABCD 2 score may be used to estimate the likelihood that a patient is having a TIA mimic. However, many have subsequently found high levels of carotid stenosis, atrial fibrillation, and other cardiac sources of embolism in patients with ABCD 2 less than 3. Based on these and other studies, the 2016 American College of Emergency Physicians (ACEP) policy for the evaluation of patients with suspected TIA in the ED recommends that physicians do not rely on current scores, such as the ABCD 2 score, to identify patients who may be discharged directly from the ED.
Risk Factor | Points |
---|---|
Age ≥ 60 | 1 |
Blood pressure (BP) | 1 |
Systolic BP ≥ 140 mm Hg or diastolic BP ≥ 90 mm Hg | |
Clinical features of transient ischemic attack (TIA) (choose 1) | |
Unilateral weakness with or without speech impairment or | 2 |
Speech impairment without unilateral weakness | 1 |
Duration | |
TIA duration ≥ 60 min | 2 |
TIA duration 10–59 min | 1 |
Diabetes | 1 |
Total ABCD 2 score | 0–7 |
Observation Unit Care
Patient selection
Several protocols have been reported ( Box 1 ). Patients are selected after being seen in the ED and given the diagnosis of TIA by an attending emergency physician. Patients are excluded from the Accelerated Diagnostic Protocol (ADP) if they have (1) any persistent acute neurologic deficit or crescendo TIAs; (2) a positive head CT; (3) a known cardioembolic source; (4) known carotid stenosis (>50%); (5) nonfocal symptoms; (6) significant other acute or chronic comorbidities necessitating inpatient admission.
Inclusion
- •
Transient ischemic attack (TIA): resolved acute deficit, not crescendo TIAs
- •
Negative head computed tomography (CT) (unless prompt MRI planned; with a normal examination and not high risk for bleed)
- •
Workup can be completed within ∼18 hours
Exclusion
- •
Head CT imaging positive for bleed, mass, or acute infarction
- •
Known extracranial embolic source: history of atrial fibrillation, cardiomyopathy, artificial heart valve, endocarditis, known mural thrombus, or recent myocardial infarction
- •
Known carotid stenosis (>50%)
- •
Any persistent acute (<72 hour) neurologic deficit or crescendo TIAs
- •
Nonfocal symptoms; for example, confusion, weakness, seizure, transient global amnesia
- •
Hypertensive encephalopathy
- •
Unable to pass emergency department dysphagia screen
- •
Severe headache or evidence of cranial arteritis
- •
Acute medical or social (poor home support) issues requiring inpatient admission
- •
Prior large stroke, making serial neurologic examinations problematic
- •
Pregnancy
Potential interventions
- •
Neuro checks every 2 hours to detect stroke, crescendo TIA, and so forth
- •
Neurology consult to detect occult stroke
- •
Fasting lipid panel, HgA1c
- •
Carotid imaging with MRI/magnetic resonance angiography (MRA) to detect infarct or surgical carotid stenosis (>50%)
- ○
If contraindications to MRI/MRA and good renal function, then CT angiography of head and neck vessels
- ○
If contraindications to MRI/MRA and poor renal function, then Doppler of neck vessels
- ○
- •
2-dimensional echocardiography as indicated by neurology to detect a cardioembolic source
- •
Cardiac monitoring for at least 12 hours for paroxysmal atrial fibrillation
- •
Antiplatelet therapy (aspirin ⇒ If already on aspirin then clopidogrel OR dipyridamole/aspirin)
- •
Stroke preventive educational materials (lipids, smoking, diabetes mellitus, hypertension, obesity, alcohol, stroke)
Disposition
Home
- •
No recurrent deficits, negative workup
- •
Clinically stable for discharge home (ie, on antiplatelet therapy and stroke-prevention medications)
- •
Hospital
- •
Recurrent symptoms/deficit
- •
Evidence of treatable vascular disease; for example, greater than 50% stenosis of neck vessels
- •
Evidence of embolic source requiring treatment (eg, heparin/coumadin); for example, mural thrombus, paroxysmal atrial fibrillation
- •
Unable to complete workup or safely discharge patient within timeframe
- •
Physician judgment
- •
Observation unit interventions
The protocol has 4 key diagnostic components: (1) carotid imaging (2) serial clinical evaluation, which includes serial neuro-assessments every 2 hours by NIHSS-trained nurses, emergency physicians, and associate providers, and a neurologist consultant; (3) cardiac monitoring for at least 12 hours; and (4) echocardiography. Because brain MRI is more sensitive for tissue injury, and generally takes longer than CT to obtain, it is obtained in the EDOU. Additionally, screening for treatable conditions that contribute to stroke occurs, such as uncontrolled diabetes, hypertension, smoking, and substance abuse. EDOU patients are often admitted as inpatients if they develop recurrent neurologic symptoms or a stroke, if significant carotid stenosis or a cardioembolic source of TIA is identified, if it is not possible to complete the workup or safely discharge the patient home within 24 hours, or if admission for other reasons is needed.
The 2009 ASA/AHA guidelines recommend that noninvasive imaging of the cervicocephalic vessels should be performed routinely as part of the evaluation of patients with suspected TIAs (Class I, Level of Evidence A). There are several options to consider for carotid imaging. Because guidelines recommend that an MRI is obtained in patients with a TIA, a magnetic resonance angiogram (MRA) of the brain and neck vessels also can be performed. If MRI cannot be performed, then CT angiography (CTA) may be obtained. If CTA cannot be obtained, then carotid Doppler imaging may be obtained. Guidelines from ACEP, the National Stroke Association, and ASA do not specify which imaging modality is superior. Recommendations may be best summarized in the ASA guidelines, which specify that imaging may be selected based on “local availability, expertise, and patient characteristics.”
Serial examinations to detect occult stroke are important. In one study, serial examinations were the diagnostic intervention that identified most patients with a positive clinical outcome leading to inpatient admission. Serial examinations may be performed by the EDOU nurses, advanced practice providers, managing emergency physician, and neurology consultants. The examination frequency should allow time to diagnose an interval stroke and initiate reperfusion therapy within 3.0 to 4.5 hours from the time last known normal. For this reason, examinations every 2 hours are often used. It is ideal for staff to be trained in a structured stroke examination, such as the NIHSS, for better consistency and to focus on the essential stroke features. It is important to focus serial examinations on the specific clinical features for which the patient presented to the ED because this represents the vascular territory that is most at risk; for example, if a patient’s presenting TIA symptom was dysarthria and left arm weakness. That specific feature should be reexamined on serial examinations, although other neurologic features may be reexamined as well.
Cardiac monitoring for paroxysmal atrial fibrillation, or other major dysrhythmias, occurs following an initial electrocardiogram (ECG) in the ED to screen for atrial fibrillation. Anticoagulation of patients with atrial fibrillation is associated with a 61% reduction in the annual risk of stroke. Patients with TIA with untreated atrial fibrillation may be admitted as inpatients. Although initial ED and observation unit cardiac monitoring for paroxysmal atrial fibrillation is important, studies of patients with TIA or cryptogenic strokes suggest that more prolonged monitoring improves detection rates in selected patients. The role of prolonged outpatient monitoring continues to be refined and is generally guided by the neurology consultant or physician who will be following the patient after discharge.
The role of echocardiography in the evaluation of patients with TIA is less clear. The 2009 ASA guidelines suggest that transthoracic echocardiography is reasonable in patients in whom no cause of TIA has been identified to detect conditions, such as patients with patent foramen ovale (PFO) or valvular disease that might benefit from specific treatments (Class IIa, Level of evidence B). However, a recent trial sheds doubt on the efficacy of PFO closure. It is also unclear which patients benefit from echocardiography during their index visit to initiate time-sensitive stroke-prevention treatments, compared with patients who may be discharged for outpatient testing.
Although screening for and initiating treatment of stroke risk factors are not associated with a decrease in the short-term stroke risk, they are associated with considerable benefit. Patients presenting to the ED with TIA are likely to represent a unique teachable moment when these interventions are most effective. For this reason, such interventions are required for hospitals to be accredited as a primary stroke center by the Joint Commission for the Accreditation of Hospital Organizations. The details of these prevention measures are often coordinated with neurology consultants or primary care physicians with whom patients will be following up.
Beneficial outcomes of transient ischemic attack accelerated diagnostic protocols
Outpatient “TIA clinics” in Europe have been shown to be effective not only in shortening time and reducing overall cost of evaluation, but actually reduce the number of subsequent strokes. Similar results have been reported in the United States using EDOUs. In a prospective randomized trial of 149 patients with TIA comparing an EDOU TIA ADP with inpatient admission with standing orders, Ross and colleagues found that ADP patients had a significantly shorter length of stay (25.6 vs 61.2 hours) and lower 90-day costs ($890 vs $1510) with no increase in adverse outcomes. ADP patients were more likely to complete necessary testing, and 85% were discharged from the EDOU. Nahab and colleagues performed a before and after validation of this ADP with 142 patients with TIA at a separate academic medical center. Findings were similar in terms of length of stay and cost differences, with no differences in compliance with stroke quality measures, such as diabetes screening, hyperlipidemia screening, or antithrombotic therapy. Others have reported similar findings as these.
These, and other issues, support the use of an ADP for patients with TIA as a best practice for selected patients relative to traditional care. The 2016 ACEP policy “Critical Issues in the Evaluation of Adult Patients with Suspected TIA in the ED,” asks the question “In adult patients with suspected TIA, can a rapid ED-based diagnostic protocol safely identify patients at short-term risk for stroke?” The patient management recommendations include that “In adult patients with suspected TIA without high-risk conditions, a rapid ED-based diagnostic protocol may be used to evaluate patients at short-term risk for stroke” (Level B recommendation). The National Stroke Association recommendations for care of patients with a TIA recognize the role of EDOUs as well.
Transient ischemic attack
Background
In the United States, stroke is the fifth leading cause of death. With roughly 800,000 strokes occurring annually, stroke is also a leading cause of disability, costing the United States roughly $34 billion per year. Eighty-five percent of strokes are ischemic, and between 15% and 30% of ischemic strokes are preceded by a transient ischemic attack (TIA). An estimated 240,000 TIAs occur annually. TIA has been redefined by the American Stroke Association (ASA) as a “(t)ransient episode of neurologic dysfunction caused by focal brain, spinal cord, or retinal ischemia without acute infarction.” Notable updates in this definition include the addition of “infarction” found on brain imaging and the removal of time elements such as 24 hours or 1 hour. Symptoms lasting more than 1 hour have a less than 1 of 6 chance of not having a stroke. Patients with clinical resolution of stroke symptoms that show brain infarction on computed tomography (CT) imaging have a fourfold higher risk of subsequent stroke, and those with a new infarct on MRI have a fivefold higher incidence of subsequent stroke.
Although patients with TIA present with no residual deficits, they are at significant risk of subsequent vascular events. Overall, their approximate 90-day stroke rate is 12%, with half of these strokes occurring within the first 1 to 2 days after a sentinel TIA. In fact, this 90-day stroke rate is more than twice that of patients who have suffered an actual stroke (4%), suggesting that patients with TIA present in a state of greater vascular vulnerability. Yet this is confounded by the fact that these patients appear clinically normal, and most will not have an adverse event during or shortly after their TIA, making TIA an ideal condition for an ED observation unit (EDOU).
Emergency Department Management
The initial ED evaluation and subsequent EDOU management of patients with TIA is focused on the early detection and prevention of stroke, the detection of alternate causes of TIA symptoms, and the prevention of future strokes. This includes a history that includes symptom duration, vascular risk factors, and other potential etiologies. Ideally, the examination should include a structured stroke examination, such as the National Institutes of Health Stroke Score (NIHSS), to improve the clinical detection of small or subtle strokes. Obtaining an electrocardiogram and placing the patient on cardiac monitoring is important for the detection of atrial fibrillation. Blood testing should include a complete blood count, serum glucose, and electrolytes. A prothrombin time is often ordered, as well, in the event that the patient subsequently develops stroke requiring reperfusion therapy. Brain imaging, with either CT or MRI, should be obtained if possible. Although the safety of delaying imaging is unknown, there are several benefits to timely imaging.
The transient ischemia that occurs in patients with TIA may be due to a number of different vascular events ( Table 1 ). Causes of TIA may be broadly classified as intracranial vascular, extracranial vascular, and cardioembolic. When an etiology is not identified, the etiology is classified as cryptogenic. Although this refers to patients with stroke, the testing for causative etiologies for TIA is similar.
Etiology | Causes |
---|---|
Intracranial vascular | Atherosclerotic Branch occlusive disease |
Extracranial vascular | Carotid plaque with arteriogenic emboli Flow-limiting stenosis |
Cardioembolic | Paroxysmal atrial fibrillation Paradoxic embolism Valve disease with clot formation Intramural clot formation associated with dyskinesis |
Cryptogenic | Cardiac etiologies: Paroxysmal atrial fibrillation, subacute bacterial endocarditis (culture negative), papillary fibroelastoma Subtle arterial dissections Central nervous system vasculitis Hypercoagulable state in setting of cancer Metabolic disorders: hyper-homocysteine, Fabry disease Plaque in aortic arch Human immunodeficiency virus, central nervous system infection Drug abuse Genetic etiologies (eg, cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy [CADASIL], mitochondrial encephalomyopathy, lactic acidosis, and stroke-like episodes [MELAS]) |
It is important to consider conditions that commonly mimic TIA. These include epilepsy, postictal paralysis, complex migraines, such as hemiplegic or ocular migraines, cervical disc disease, transient global amnesia, intracranial lesions, such as subdural hematomas and mass lesions, inner ear disease with vertigo, and metabolic derangement of the serum glucose, sodium, or calcium.
A unique challenge with the diagnosis of TIA is the lack of a clear gold standard for this diagnosis. When a neurologist is used as the gold standard, false-positive diagnostic rates range from 25% to 38%. There is even poor agreement between fellowship-trained stroke neurologists.
To improve diagnostic performance, TIA risk stratification scores have been developed. The most common is the ABCD 2 score ( Table 2 ). American Heart Association (AHA)/ASA guidelines recommend admission or discharge based on a total score greater or less than 3. Like other clinical prediction tools, the ABCD 2 score may be used to estimate the likelihood that a patient is having a TIA mimic. However, many have subsequently found high levels of carotid stenosis, atrial fibrillation, and other cardiac sources of embolism in patients with ABCD 2 less than 3. Based on these and other studies, the 2016 American College of Emergency Physicians (ACEP) policy for the evaluation of patients with suspected TIA in the ED recommends that physicians do not rely on current scores, such as the ABCD 2 score, to identify patients who may be discharged directly from the ED.
Risk Factor | Points |
---|---|
Age ≥ 60 | 1 |
Blood pressure (BP) | 1 |
Systolic BP ≥ 140 mm Hg or diastolic BP ≥ 90 mm Hg | |
Clinical features of transient ischemic attack (TIA) (choose 1) | |
Unilateral weakness with or without speech impairment or | 2 |
Speech impairment without unilateral weakness | 1 |
Duration | |
TIA duration ≥ 60 min | 2 |
TIA duration 10–59 min | 1 |
Diabetes | 1 |
Total ABCD 2 score | 0–7 |
Observation Unit Care
Patient selection
Several protocols have been reported ( Box 1 ). Patients are selected after being seen in the ED and given the diagnosis of TIA by an attending emergency physician. Patients are excluded from the Accelerated Diagnostic Protocol (ADP) if they have (1) any persistent acute neurologic deficit or crescendo TIAs; (2) a positive head CT; (3) a known cardioembolic source; (4) known carotid stenosis (>50%); (5) nonfocal symptoms; (6) significant other acute or chronic comorbidities necessitating inpatient admission.
Inclusion
- •
Transient ischemic attack (TIA): resolved acute deficit, not crescendo TIAs
- •
Negative head computed tomography (CT) (unless prompt MRI planned; with a normal examination and not high risk for bleed)
- •
Workup can be completed within ∼18 hours
Exclusion
- •
Head CT imaging positive for bleed, mass, or acute infarction
- •
Known extracranial embolic source: history of atrial fibrillation, cardiomyopathy, artificial heart valve, endocarditis, known mural thrombus, or recent myocardial infarction
- •
Known carotid stenosis (>50%)
- •
Any persistent acute (<72 hour) neurologic deficit or crescendo TIAs
- •
Nonfocal symptoms; for example, confusion, weakness, seizure, transient global amnesia
- •
Hypertensive encephalopathy
- •
Unable to pass emergency department dysphagia screen
- •
Severe headache or evidence of cranial arteritis
- •
Acute medical or social (poor home support) issues requiring inpatient admission
- •
Prior large stroke, making serial neurologic examinations problematic
- •
Pregnancy
Potential interventions
- •
Neuro checks every 2 hours to detect stroke, crescendo TIA, and so forth
- •
Neurology consult to detect occult stroke
- •
Fasting lipid panel, HgA1c
- •
Carotid imaging with MRI/magnetic resonance angiography (MRA) to detect infarct or surgical carotid stenosis (>50%)
- ○
If contraindications to MRI/MRA and good renal function, then CT angiography of head and neck vessels
- ○
If contraindications to MRI/MRA and poor renal function, then Doppler of neck vessels
- ○
- •
2-dimensional echocardiography as indicated by neurology to detect a cardioembolic source
- •
Cardiac monitoring for at least 12 hours for paroxysmal atrial fibrillation
- •
Antiplatelet therapy (aspirin ⇒ If already on aspirin then clopidogrel OR dipyridamole/aspirin)
- •
Stroke preventive educational materials (lipids, smoking, diabetes mellitus, hypertension, obesity, alcohol, stroke)
Disposition
Home
- •
No recurrent deficits, negative workup
- •
Clinically stable for discharge home (ie, on antiplatelet therapy and stroke-prevention medications)
- •
Hospital
- •
Recurrent symptoms/deficit
- •
Evidence of treatable vascular disease; for example, greater than 50% stenosis of neck vessels
- •
Evidence of embolic source requiring treatment (eg, heparin/coumadin); for example, mural thrombus, paroxysmal atrial fibrillation
- •
Unable to complete workup or safely discharge patient within timeframe
- •
Physician judgment
- •