Treatment of Seizures in the Emergency Department: Pearls and Pitfalls
- Dec 17th, 2015
- Brit Long
Author: Raul Hernandez, MD (EM Resident Physician, SUNY Downstate/Kings County Hospital) and Mark Silverberg, MD (EM Attending Physician and Associate Residency Director, SUNY Downstate/Kings County Hospital) // Edited by: Alex Koyfman, MD (@EMHighAK, EM Attending Physician, UT Southwestern Medical Center / Parkland Memorial Hospital) and Brit Long, MD (@long_brit, EM Chief Resident at SAUSHEC, USAF)
Seizure-related complaints comprise approximately 1% of all emergency department visits.  This includes patients with first-time seizures and those with an established history of epilepsy. The difference between epilepsy and initial seizures will be discussed in the next section. Acute and chronic causes exist as triggers of seizures and include acute intracranial hemorrhage, cerebral injuries/scars, tumors, infections, metabolic derangements, and, of course, congenital causes of epilepsy.  Increasing age also confers greater likelihood of seizures or electroencephalographic abnormalities. [3, 4]
Defining Seizures and Epilepsy
The timing of seizures in relation to triggers or the presence of a discernible trigger is the basis for their classification as provoked or unprovoked. Provoked seizures are those that occur within 7 days of an acute neurologic, systemic, metabolic, or toxic (e.g., visual, intoxication or detoxication) insult. Unprovoked seizures are either farther out than 7 days from any precipitating factor or are not associated with any discernible cause. 
In addition, not all seizures are of the generalized convulsive type. Other types of seizures may involve focality. That is they may display sensory and/or motor effects in only one area of the brain. These other types are absence, simple partial, and complex partial seizures. As you may already be aware, the difference between a simple and complex partial seizure is that alteration in consciousness occurs only with complex seizures. Consideration must be given to the development or presence of status epilepticus in patients presenting with the complaint of seizure regardless of the subtype. 
Status epilepticus has historically had many differing definitions. The American College of Emergency Physicians (ACEP) 2014 clinical policy on patients presenting to the emergency department with seizures define status epilepticus as either unremitting convulsive seizure activity lasting 20 minutes or more, or intermittent seizures without regaining full consciousness in between. They also state that a generalized convulsive seizure lasting for at least 5 minutes can been defined as constituting status epilepticus.  This continuous seizure activity can change, evolving from a convulsive state to a subclinical state due to fatigue or exhaustion of muscles.
As the name suggests, this type of status epilepticus is more difficult to diagnose since the most obvious part of the seizure — the convulsion — has ceased. This is especially true in patients already comatose upon arrival. Unfortunately, subclinical seizures also have a much worse prognosis than other types.  Diagnosing this type of seizure activity requires EEG analysis, a topic we will further explore later.
Finding the Cause
Rather than a pathophysiologic explanation of seizures, the focus of this section will be diagnostic strategies for the patient presenting with seizures. The best strategy, by far, is a good history and physical examination since up to fifty percent of patients who present with a first-time seizure are actually found to have had similar preceding episodes and can possibly be diagnosed with epilepsy. 
As mentioned previously, there are provoked and unprovoked seizures. In clinical practice it is often difficult to tell the two apart. However, sometimes provoked seizures are secondary to disease states with clinically overt signs and symptoms such as intoxication, withdrawal (ethanol, benzodiazepines), trauma, meningitis, psychiatric issues, and metabolic derangements (e.g., hypoglycemia). This is by no means an exhaustive list, and coming up with one would be difficult as there are numerous causes for seizures. Often the history and/or physical exam help raise suspicion for the cause of the seizure (e.g., fever, stiff neck and altered mental status preceding a seizure should raise suspicion for meningitis).
The American Academy of Neurology (AAN) states in their practice parameters that the diagnostic and therapeutic modalities for provoked and unprovoked seizures are different in that addressing provoked seizures appropriately means providing intervention for the acute process causing them, whereas with unprovoked seizure treatment is aimed at seizure control. 
Provoked Seizure, Back to Baseline
According to the ACEP clinical policy, patients who have had a first-time provoked seizure, are back to baseline, and are otherwise asymptomatic should be evaluated and treated appropriately for the cause of their seizure. Anti-epileptic medication, which in this paper are defined as preventive (e.g., levetiracetam) and not abortive (e.g., lorazepam), need not be initiated.  The underlying factor or factors that caused the seizure should be addressed with the understanding that this should remove the stimulus that caused the seizures in the first place. [4, 6]
First-time Provoked Seizure, Not Back to Baseline
Once again, the offending factor or factors need to be addressed in this population. This subset of patients is at higher risk of having seizure recurrence.  These patients represent a subset that could benefit from additional testing such as electroencephalography in addition to more standard labs and imaging. EEG may be normal in up to fifty percent of patients with an actual seizure. However, for some patients it may show important information such as the presence of status epilepticus that is not overt.  Which labs and whether to image should be determined by the suspected causative factor. For instance, in a patient with a history of cancer who has experienced a seizure without return to baseline, a computer tomographic scan of the head may show metastases or some other intracranial process.  Remember that alcohol and its lack (withdrawal) is considered a risk factor for seizures. [4, 9]
First-time Unprovoked Seizure, Back to Baseline
Though these adult patients have returned to mental baseline, they will still require testing for glucose and sodium levels, a pregnancy test in women of childbearing age, and either a computed tomography (CT) head scan or a magnetic resonance imaging (MRI) study of the brain. Initiating an anti-epileptic medication is left to the discretion of the physician by ACEP, but if it is a true unprovoked first-time seizure then no medication is necessary. This is because one seizure does not constitute the diagnosis of epilepsy, and early initiation of therapy does not seem to provide protection from development of future seizures in the long-term. If a history of stroke or other central nervous system insult can be elicited, it changes management drastically. The number needed to treat (NNT) for prevention of one additional seizure with antiepileptics in this setting is five. It is estimated that up to 50% of patients presenting for first-time seizures have had previous similar incidents, which at the current presentation would give very high suspicion for a diagnosis of epilepsy. If such history can be elicited, then it makes treatment after recurrent seizure presentation warranted. 
Regardless of history of previous CNS insult, CT of the head is an important test in these patients, as this study alone has been shown to change management in approximately ten percent of cases. It is unclear whether MRI is a better study than CT in these cases, and it is unclear whether it is strictly necessary to do imaging right away or if it can wait for an outpatient appointment. [4, 6] Of course, social circumstances must be taken into account when deciding whether to scan at presentation or have imaging done on an outpatient basis. [1-3, 9]
Short-term risk is not well-defined in this patient population. Most studies look at risk of repeat seizure in months or years but not days. Unfortunately no prospectively derived guidance exists, although an algorithm has been proposed, and the College of Emergency Medicine (United Kingdom) practice guidelines published in 2009 reflect this. They propose that a patient can be discharged to home after a first generalized seizure if the following all exist: normal glucose, CT head scan, electrocardiogram, renal function tests, electrolytes including calcium, and complete blood count. In addition, the patient must have no abnormal neurologic signs or symptoms after a full recovery, normal vital signs, good social support including established follow-up and good likelihood of the appointment being kept, and a responsible adult to watch the patient at home. [7, 8] The AAN practice parameters discuss the uncertainty of the value of an electroencephalogram in this group of patients but do recommend it as a valuable study, especially when done within the first 24 hours after a seizure. 
Repeat Unprovoked Seizure/History of Epilepsy, Back to Baseline
If this seizure is similar to previous seizures, then these patients require evaluation of their anti-epileptic drug levels (if measurable) and assessment of other triggering factors. Often patients will be controlled on their antiepileptics regimen even at subtherapeutic levels but will have other triggers (e.g., lack of sleep, changes in or new medications, vomiting) which cause breakthrough seizures.  If they have been non-compliant with their AEDs medications, then they can be given a loading dose. There are many different medications and each has its own dosing regimen. Thankfully ACEP clinical policy does include a table with many of these agents.  Of note, intravenous phenytoin and, to a lesser degree, IV fosphenytoin, are more likely to cause hypotension without any proven benefit in decreasing seizure recurrence.
If the presenting seizure is different from those in the past then a more comprehensive work up is warranted looking for inciting factors. Once triggering factors and medication compliance have been addressed these patients may be able to go home. 
First-time or Repeat Unprovoked Seizure, Not Back to Baseline
This is a patient with a truly unprovoked seizure that has either persistent altered mental status or continued seizure activity. In addition to a work-up including consideration of lumbar puncture, imaging, electrocardiogram, and labs as described above (first-time unprovoked seizures, back to baseline), these patients may warrant electroencephalography (EEG) if it is available. [2,3] Status epilepticus, which is discussed next, must be considered in addition to any underlying factors that may actually make this a provoked seizure.
Status epilepticus (SE) is truly a medical emergency. This phenomenon can present secondary to absence, simple partial, complex partial, and generalized convulsive seizures. There are two types: convulsive and non-convulsive status epilepticus (NCSE). Non-convulsive can evolve from all four types of seizures. It can also sometimes be mistaken for a coma.  The most dangerous type of SE is termed subtle SE, a name referring to its lack of convulsive activity with continued seizure activity in the brain. SE develops from a generalized convulsive epileptic seizure. It can sometimes be refractory to therapy and may require intubation and placing the patient in a medicine induced coma. [5, 6, 13] Treatment strategies are described below but as for disposition, this should be carefully discussed with the patient’s own neurologist or primary care physician. Admission is normally warranted. 
Clinical Findings and the Electroencephalogram
The physical exam can reveal information that can raise suspicion for true seizure activity and also status epilepticus. Abnormal neurological examination, predisposing history (<6 months of age, closed head injury, recent cerebrospinal fluid shunt revision, malignancy, neurocutaneous disorder, focal onset of seizure, no history of alcohol abuse, history of cysticercosis, altered mental status, age greater than 65, and seizure lasting >15 minutes) are probably predictive of an abnormal CT study in this context.  Another study also found some clinical features that could be predictive of status epilepticus on physical exam. These are remote factors for a seizure (stroke, tumor, previous neurosurgery, dementia, and meningitis), GCS <15, and abnormal ocular movements (nystagmus, hippus, or sustained eye deviation in any direction). The individual sensitivity and specificity of each of these findings was low. However, the study authors report a combined sensitivity for “remote risk factors for seizures” and “eye movement abnormalities” of 100%. If the sensitivity truly is that high, it may provide a good screening test for which patients with altered mental status should receive an urgent EEG. 
Alcohol abuse should also prompt a physician to have higher suspicion for SE. For instance, one review found that alcohol abuse was related to the presentation of seizures in 20-40% of patients presenting to an emergency department. It also found that among patients with SE, up to 20% are likely withdrawing from alcohol.  A separate study by Zehtabchi et al. found that in ED patients with undifferentiated altered mental status, EEG abnormalities existed in 78% of subjects, and NCSE was detected in five percent. 
CT scan is the most common brain imaging study performed on patients with first time seizures. A review article by the American Academy of Neurologists found treatment changed due to CT findings in 9-17% of patients. This review also found that patients with AIDS and a first seizure were commonly diagnosed with abnormalities via CT. Toxoplasmosis was a frequent finding. 
Although neurologists tend to prefer MRI over CT, there is no published data that shows superiority of MRI over CT. The practice parameters published by the AAN also state that CT is an adequate first brain imaging study for a patient with first-time seizure and even in some cases of repeat seizures. [4, 6]
Although we tend to think of the brain first when we think of seizures, we should also be thinking about the heart. Surprisingly, it has been known for quite some time that seizures can cause dysautonomia leading to cardiac dysrhythmias. [11, 12] Equally important is remembering that seizures can be caused by ingestion of substances that affect the fast sodium channels such as tricyclic antidepressants and other toxins that may also affect the heart. [2, 6] An ECG is appropriate as a screening tool for seizure patients presenting to the emergency department. [6-8]
Lumbar puncture should not be done on every patient presenting to the ED with a seizure, especially if they are back to baseline. It should instead be done selectively on patients with persistently altered mental status, fever, meningismus, severe headache, localized neurological signs and/or symptoms, and immunosuppression. [4,6]
A presentation of seizures during pregnancy should prompt a screening for eclampsia. The clinician should also ascertain if a history of epilepsy already exists for the patient or if it is a first-time provoked or unprovoked seizure due to the one of etiologies previously described. Most women do not experience an increase in seizures during pregnancy, but for those who do, it might be attributable to poor medication compliance (that can be exacerbated by vomiting related to pregnancy), inappropriate reduction of AEDs, and/or a decreased amount of sleep. [6, 15]
The same management strategies apply to pregnant patients, especially when the patient is experiencing status epilepticus. Treatment should proceed taking into account two things. First, during pregnancy measurable AED levels in the serum may decrease secondary to a physiologically predicted increase in volume of distribution, hepatic clearance, and renal clearance. Be aware that although measurable drug levels may be lower there may also be a higher level of free-drug levels due to the decrease in plasma proteins that also accompany the increased plasma volume. The second thing to know is that valproate should be avoided in pregnancy, as this AED has been shown to have a high rate of teratogenicity, especially in combination with lamotrigine. 
It is important that patients understand the risk of teratogenicity with AEDs. Therapy with one AED during pregnancy confers a three percent risk of fetal malformation, which is slightly above the background risk without AEDs. Therapy with two or more AEDs increases that risk to approximately 17%. Most severe malformations occur if the fetus is exposed in the early stages (first-trimester) to AEDs, and the risks associated with SE to both the mother and the fetus far outweigh the risks associated with AED therapy. [6,15]
Before starting the discussion of what seizure abortive medications can be given to a patient experiencing seizures, it is important to recall that some anti-epileptic medications (e.g., phenytoin, carbamazepine) can actually induce seizures in some patients. It is important to gather information about recent changes or new medications. If seizures began in your patient after administration of a new or increased dose of AED, then withdrawal of this medication is advisable followed by treatments described below to break any persistent seizure activity. 
Regardless of type of seizure, patients actively seizing should receive benzodiazepines according to both emergency medicine and neurology literature. [2, 5, 6, 13, 14] Please keep in mind that patients may suffer from hypotension and decreased respiratory drive when administered these medications. Cardiopulmonary monitoring should be in place. Prasad et. al compared many medications commonly given during status epilepticus and found that in the ED setting, lorazepam given IV decreases the chance that SE will continue and the chance of needing more medications to stop SE. They also found midazolam was “probably better” than lorazepam for pre-hospital treatment because it can be given IM, obviating the need for IV access, and confers decreased risk of hospitalization and intensive care unit admissions when administered in the field. Surprisingly, they also found some data suggesting levetiracetam IV works equally well as lorazepam IV at stopping seizures in the in-hospital setting with less propensity for causing issues such as hypotension and need for mechanical ventilation. However, because it was only found in one small study, the authors continue to recommend lorazepam IV as the initial treatment of choice over other benzodiazepines or phenytoin.  Because not all drugs are available at all times and because IV is not always a feasible route of administration, suggested dosing for three benzodiazepines are as follows [5, 6]:
- lorazepam 4 mg IV push over 2 minutes (may be repeated in 5-10 min if seizure activity continues)
- diazepam 10 mg IV or rectal (may be repeated as above)
- midazolam 10 mg IV (may also be given IM if no IV access)
If benzodiazepines are ineffective at stopping seizure activity, a more nuanced approach is advocated by the AAN practice parameters. If the patient has a history of absence seizures and is in SE due to this phenomenon then bolus administration of valproic acid (25-45 mg/kg at a rate of 6 mg/kg/min) or phenobarbital (20 mg/kg at 50 mg/kg/min) is recommended. If the patient has a history of frontal or temporal lobe epilepsy and is in partial SE they may require phenytoin (15-18 mg/kg at 25-50 mg/min; rate not to exceed 50 mg/min) or an equivalent dose of fosphenytoin (at 150 mg/min). [5, 6] Note that phenytoin can cause hypotension when given IV. Fosphenytoin also causes this side effect but at a much lower rate. Patients being administered either medication should be on cardiopulmonary monitoring. Fosphenytoin is also safe to give IM, whereas phenytoin is very toxic to vascular walls and causes necrosis if it extravasates into surrounding tissue during IV administration. This is why phenytoin should only be administered IV when a large vein is utilized. Even so, limb edema, ischemia, and discoloration can still occur. For these reasons and the rapidity with which it can be infused, fosphenytoin is often preferred over phenytoin. 
The third-line medications are very powerful and have side effects that are commensurate to their potency. These medications are anesthetics being dosed at induction levels, and intubation followed by mechanical ventilation will be necessary along with appropriate cardiopulmonary monitoring.  If the patient is suffering from subtle SE, then administer induction dose anesthetics as described below without delay. Success of GABAergic medications (i.e., benzodiazepines) in this patient population is compromised due to the modification of GABA A receptors from the continuous seizure activity, and a high failure rate of these medications should be anticipated. 
If patients with SPSE and CPSE do not respond to previously mentioned treatment, the next medications in the algorithm are phenobarbital 20 mg/kg or valproic acid at 25-45 mg/kg. The authors warn that some neurologists are reluctant to go beyond this point to medications such as propofol because they carry significant morbidity and mortality risks. However, they identify a population of young patients with no other comorbidities who may benefit from treatment with midazolam (0.2 mg/kg bolus and 0.1-0.4 mg/kg/hr infusion), propofol (2 mg/kg bolus and 5-10 mg/kg/hr infusion), and thiopental (2-3 mg/kg bolus and 3-5 mg/kg/hr infusion). They make these recommendations in acknowledgement of a greater than 50% recurrence rate of seizure activity in this population. [5, 6]
Claasen et al.  compared two of the medications mentioned above and one other: midazolam, propofol, and pentobarbital in the setting of refractory SE. They found that titration of these medications to EEG background (isoelectric or burst) suppression rather than simply suppression of seizures led to a lower frequency of breakthrough seizures with a higher rate of hypotension also occurring. The study excluded simple partial and absence SE. Unfortunately, no mortality benefit was found with these medications even if the titration goal of EEG background suppression was met.
Admit vs. Discharge
As discussed in the section “First-Time Unprovoked Seizure, Back to Baseline” above, there are some guidelines, albeit not prospectively tested, for who can be considered discharged to home. To recapitulate, patients who experienced an unprovoked seizure but are otherwise healthy, have normal vital signs, a normal glucose level, CT head scan, electrocardiogram, renal function tests, electrolytes including calcium, and a complete blood count may be considered for discharge. This only applies to patients without abnormal neurologic signs or symptoms who have fully recovered from their seizure. Good social support is important as well because the patient will need a responsible adult to watch them at home. Prior to the patient leaving the emergency room, follow-up should be secured for the patient and the physician should feel there is a good likelihood of the patient getting to that appointment. [7, 8]
Pitfalls in the Treatment of Seizures
- Forgetting your ABCs: if a patient is continually seizing they may not be able to protect their airway or ventilate properly. If intubation is required to properly oxygenate, there should be no delay of intubation. Cardiopulmonary monitoring should also occur, especially in light of the rare but possibly fatal dysautonomia that can accompany seizures. 
- Using long-acting paralytics can mask seizure activity and if one has been used then an EEG might be considered to monitor the patient for seizure activity. 
- The most common finding on ECG is sinus tachycardia, but more concerning arrhythmias can occur. An ECG can be performed once a patient is not convulsing to check for these arrhythmias. 
- Forgetting to do a fingerstick glucose test. This is a relatively inexpensive test that may save a lot of work-up, administration of unnecessary medications, and provides a simple solution (50 mg of dextrose preceded by thiamine if suspicion for malnutrition exists). 
- Just because a patient is not convulsing does not mean seizure activity has ended. Non-convulsive status epilepticus exists and should be suspected until ruled out. [1-7, 10] 5 minutes of continual seizure activity define status epilepticus in the clinical setting because damage to brain cells can occur within this short time span, even in the absence of convulsions. 
- Phenytoin/fosphenytoin may not appropriately treat toxin-induced seizures because these AEDs work on sodium channels, not GABA receptors. This is most important to remember with alcoholics, who have the highest early seizure recurrence. Other medications that can cause the same type of resistance during a detoxication phase are benzodiazepines. 
- Do not assume the seizure is solely due to low levels of AEDs which can be measured. Instead seizures might be due to other stressors (lack of sleep, metabolic derangements, vomiting in pregnant women, etc.). Though if measured levels are low consideration of a loading dose is appropriate. [2, 5]
- Though there is no strong data to support or refute the practice, a lumbar puncture on immunocompromised patients can find occult intracranial infection. [4, 6]
- Know if your state requires mandatory reporting by physicians of patients with seizures to the DMV. It is important to stay compliant with this law if it exists where you practice. 
- American College of Emergency Physicians. Clinical policy: critical issues the evaluation and management of adult patients presenting to the emergency department with seizures. Ann Emerg Med. 2004;43:605-625.
- Huff JS, Melnick ER, Tomaszewski CA, et al. Clinical policy: critical issues in the evaluation and management of adult patients presenting to the emergency department with seizures. Ann Emerg Med. 2014;63(4):437-447.
- Zehtabchi S, Abdel Baki SG, Omurtag A, et al. Prevalence of non-convulsive seizure and other electroencephalographic abnormalities in ED patients with altered mental status. Am J Emerg Med. 2013;31(11):1578-1582.
- Krumholz A, Wiebe S, Gonseth G, et al. Quality Standards Subcommittee of the American Academy of Neurology; American Epilepsy Society. Practice Parameter: evaluating an apparent unprovoked first seizure in adults (an evidence-based review): report of the Quality Standards Subcommittee of the American Academy of Neurology and the American Epilepsy Society. Neurology 2007 Nov 20;69(21): 1996-2007.
- Meierkord H, Holtkamp M. Non-convulsive status epilepticus in adults: clinical forms and treatment. Lancet Neurol. 2007;6(4):329-339.
- Teran F, Harper-Kirksey K, Jagoda A. Clinical decision making in seizures and status epilepticus. Emerg Med Pract. 2015 Jan;17(1):1-24; quiz 24-5.
- Turner S, Benger J, for the College of Emergency Medicine (UK). Guideline for the management of first seizure in the emergency department. The Guidelines in Emergency Medicine Network (GEMNet). December 2009.
- Dunn MJ, Breen DP, Davenport RJ, Gray AJ. Early management of adults with an uncomplicated first generalised seizure.
- Harden CL, Huff JS, Schwartz TH, et al. Reassessment: neuroimaging in the emergency patient presenting with seizure (an evidence-based review): report of the Therapeutics and Technology Assessment Subcommittee of the American Academy of Neurology. Neurology. 2007;69(18):1772-1780.
- Husain AM, Horn GJ, Jacobson MP. Non-convulsive status epilepticus: usefulness of clinical features in selecting patients for urgent EEG. J Neurol Neurosurg Psychiatry 2003; 74: 189–91.
- Allana SS, Ahmed HN, Shah K, Kelly AF. Ictal bradycardia and atrioventricular block: a cardiac manifestation of epilepsy. Oxf Med Case Reports. 2014 May 19;2014(2):33-5. doi: 10.1093/omcr/omu015. eCollection 2014 May.
- Britton JW, Ghearing GR, Benarroch EE, Cascino GD. The ictal bradycardia syndrome: localization and lateralization. Epilepsia. 2006 Apr;47(4):737-44.
- Prasad M, Krishnan PR, Sequeira R, Al-Roomi K. Anticonvulsant therapy for status epilepticus. Cochrane Database Syst Rev. 2014 Sep 10;9:CD003723.
- Claassen J, Hirsch LJ, Emerson RG, et al. Treatment of refractory status epilepticus with pentobarbital, propofol, or midazolam: a systematic review. Epilepsia. 2002;43(2):146-153.
- Crawford P. Best practice guidelines for the management of women with epilepsy. Epilepsia. 2005;46 Suppl 9:117-24.