1. Description of the problem

Status epilepticus is the presence of seizure activity lasting for more than five minutes with impared neurologic function, or recurrent seizures without a return to baseline mental status between seizures. Seizures are a common problem in the ICU. Several independent studies have found seizure rates in the 10-20% range for hospitalized patients, with generally higher rates in critical illness settings. In addition to seizures being more common than previously thought, most of the seizure activity is subclinical; that is, there are no overt clinical signs of the underlying seizure activity except for depressed or altered levels of consciousness. The diagnosis must then be made through evaluation with electroencephalography.

In critical illness from either medical, surgical or traumatic causes patients are at an increased risk for seizure activity. Only episodic eye deviation and an altered or depressed level of consciousness have been found to be reasonable clinical indicators of underlying seizures. Generalized seizure activity occurs in some cases and is generally easily identified and treated quickly in the acute care setting. However, following a seizure if the patient does not return to baseline level of consciousness, underlying ongoing seizures should be considered. Other risk factors for seizures include a prior history of seizures and noncompliance with antiseizure medications.

2. Emergency Management

The priority for management is stopping the seizure(s). In cases of multiple recurrent seizures, early use of antiseizure medications is appropriate. In some cases aggressive intervention includes a continuous infusion of benzodiazepine or other anesthetic agents and endotracheal intubation. This is because the longer a patient remains in status, the more difficult it is to stop the seizures. As a result, rapid seizure control is an early goal of therapy.

Continue Reading

In a patient presenting in status due to a seizure lasting more than five minutes, immediate administration of intravenous (IV) benzodiazepine is recommended. Lorazepam (4-8 mg, IV) has been shown to be effective in stopping seizures in over 60% of cases. An additional dose of 4 mg may be repeated once, but if a second dose is required a loading dose of an anticonvulsant should also be considered and airway management should also be considered. Traditionally, Dilantin 15 mg/kg loading dose is given as a first-line anticonvulsant agent, though some trials of Depakote 15-20 mg/kg IV and Keppra 2000 mg IV have also been shown to be effective.

In patients without a prior history of seizure, with continued suppressed level of consciousness or confusion, or in whom new neurological deficits are suspected, immediate imaging with a non-contrast computed tomography (CT) scan should be obtained. If no cause for the seizures can be found with imaging, sampling of spinal fluid and initiation of electroencephalography are appropriate.

3. Diagnosis

The absence of acute brain injury does not reduce the risk for underlying seizures as a cause for altered or depressed levels of consciousness in a critically ill patient. Indeed, in both general medical ICUs as well as mixed medical-surgical ICUs, the reported incidence of underlying seizures has ranged from 10% to 30%. There are no definitive clinical signs or symptoms of seizures in critically ill patients. While generalized tonic-clonic seizures have a characteristic clinical appearance and are easily diagnosed, these represent the minority of seizures in the ICU. It has been previously reported that episodic eye deviation can be suggestive of ongoing seizures. In most cases, level of consciousness is reduced during and for a variable period after the seizures. However, some patients may not lose consciousness and may only have a mild motor manifestation of their seizures. Given the numerous other reasons for a depressed level of consciousness in critically ill patients, physical exam findings cannot be used to exclude seizures in a given patient. As a result, the gold standard for diagnosis of seizures is EEG.

EEG measures brain electrical activity using metallic disc or needle electrodes placed at the scalp. A conductive paste or glue is used to secure the leads to the scalp. The EEG leads are placed using an international standard approach in which even spacing is achieved by measuring the head and calculating the appropriate distances to create inter-electrode distances that are 10% lateral and 20% anterior to posterior. Typically EEG recording is continued for 24 hours in the ICU setting as patients can have recurrent seizures that would be missed if the standard EEG duration of 20 minutes were used. The EEG data must then be reviewed by a physician trained in EEG interpretation and optimally by one with experience in reading ICU patient studies.

If the patient’s level of consciousness remains altered and the EEG does not reveal seizures as the cause, other potential causes should be considered and evaluated. The EEG, even when seizures are not seen, can be very helpful in pointing the team to an underlying cause for the patient’s condition. EEG features such as triphasic waves and diffuse symmetric slowing can indicate a metabolic encephalopathy. More focal changes such as focal slowing or periodic, lateralized epileptiform discharges (PLEDs) are often seen in the setting of structural abnormalities. Periodic sharp waves in a focal region can also indicate an area of abnormal excitability suggestive of a focal lesion or other disease processes such as CJD. These results can help guide the decision to repeat imaging with either CT or magnetic resonance imaging (MRI) to better evaluate for possible structural lesions.

In the absence of imaging data to indicate a cause for seizures, sampling of spinal fluid is mandatory. Typically, 10-20 cc’s of fluid are collected and the fluid is analyzed for the number of white blood cells, number of red blood cells in the first and last tube used to collect the samples, as well as special stains to look for fungus and yeast as well as other pathogens (prions and viral particles, for example). A rare cause of uncontrolled seizures in patients who do not have a prior history of seizures is an autoimmune disorder. Testing of serum and/or spinal fluid for the presence of these antibodies can provide a diagnosis and indicate the need for plasmapheresis or IVIG administration.

4. Specific Treatment

The goal of treatment is to stop the seizure activity as quickly as possible. For clinically apparent seizures, often treatment with lorazepam or another benzodiazepine has already been initiated prior to arrival in the ICU. However, if seizures begin in the ICU, it is important to remember the early role of benzodiazepines in seizure management. When seizures continue, loading with an antiepileptic agent is indicated. More important than which agent is used, the ability to get the agent administered quickly is paramount, making the intravenously administered medications preferred agents for managing status.

A list of drugs is provided here for reference. While this represents the preferred sequence of administration at many institutions, there are insufficient data to support one series of administrations over another. What is important is that a protocol exists and is followed to ensure that appropriate care and management are provided in a quick and efficient manner. In cases in which seizures continue after two agents have been loaded, it is appropriate to begin seizure suppression doses of intravenous anesthetic agents instead of or in addition to adding a third antiseizure agent.

Drugs and dosages

— Lorazepam 4 mg IV, may be repeated once, but airway and sedation effects will need to be managed.

— Dilantin (fosphenytoin) 15-20 mg/kg IV load over 15-20 min (150 mg/min) (monitor for hypotension and arrhythmia, slow infusion rate if encountered).

— Leveteracitam 2000 mg IV load, then 1000 mg IV every 8 hours.

— Depakote (valproic acid) 15-20 mg/kg IV load over 15-20min (can be a substitute for Dilatin as a first-line agent; monitor for liver toxicity).

— Midazolam infusion with 10-mg bolus then run at 5 mg/hr and titrate to seizure suppression.

— Propofol infusion with 2 mg/kg load, then run at 2-10 mg/kg/hr and titrate to seizure suppression.

— Pentobarbital infusion with 5-10 mg/kg load, then run at 1-5 mg/kg/hr and titrate to seizure suppression.

Refractory cases

Most patients presenting in status will respond to a single agent. However, in some cases, particularly in the ICU, the seizures can remain highly refractory to treatment. These patients should be moved to the ICU if not already there, and the patient should be intubated and placed on anesthetics to suppress the seizures. EEG monitoring should be initiated as soon as possible to guide management and to confirm seizures have been controlled.

Once the seizures are stopped they should be suppressed for a full 24 hours prior to weaning any of the medications. If seizures recur with weaning of the infusions, another agent should be added. Data exist for Lyrica, Topamax and Tegretol in this setting, with growing acceptance of the newest agent Lacosamide as well. Again, seizures should be suppressed for a full 24 hours prior to trying to wean the intravenous drips again.

5. Disease monitoring, follow-up and disposition

Expected response to treatment

Patients with a single generalized tonic-clonic event often do not need more than close monitoring for subsequent events. In the ICU setting, more often patients have few to no clinical signs of seizures. In this setting, EEG monitoring is mandatory. Continuous EEG monitoring will ensure the seizures are suppressed and that they do not recur as infusions are weaned. Most patients will respond to a round or two of seizure suppression and recover to baseline. However, the level of recovery is tied to the underlying cause for the seizures. For example, large intracranial hemorrhage will produce clinical deficits unrelated to the seizures produced by the hemorrhage. In a rare few patients seizures will be extremely refractory and require prolonged periods (weeks to months) of deep sedation with anesthetic drugs, often in combinations and at high doses, to suppress the seizures. Prognosis for these refractory cases is highly variable, but includes some very good outcomes with return to work and normal function after weeks in the ICU.

As with the emergency evaluation, as soon as the patient is stable enough, imaging should be performed. MRI with contrast using a seizure protocol is the preferred study to look for an underlying cause for the seizures. Once a mass lesion or other cause for elevated intracranial pressure has been excluded, lumbar puncture and evaluation of spinal fluid for infectious causes of seizures is appropriate. Current medications should always be reviewed, as several common outpatient medications can lower seizure thresholds. Full metabolic evaluation is also warranted.


Typically the patient will remain on the antiepileptics started in the ICU for the duration of the hospitalization. Outpatient follow-up with a neurologist is recommended so that the antiseizure agents can be weaned over time in a controlled manner by an experienced provider.


The true underlying cause for seizure initiation is unknown. There are multiple clinical conditions that can lead to spontaneous seizures. These can range from structural conditions such as stroke, traumatic brain injury and surgical procedures, to infectious causes, to medications, to metabolic causes such as renal failure, drug withdrawal or overdose, and electrolyte imbalance. Even autoantibodies to select ion channels can lead to seizure activity. The mechanisms by which these perturbations cause seizure activity in the brain to develop and be sustained are far from clear and are likely to be varied in mechanism.

There are good data for some seizure causes, one being alcohol withdrawal. Ethanol binds to the GABAA inhibitory neurotransmitter receptors, which causes suppression of activity within the brain and downregulation of the receptors. As the alcohol level falls during withdrawal, these receptors become less activated and are no longer present in sufficient numbers to control the increasing levels of brain activity. In this disinhibited state of excitability, seizure thresholds are lower and spontaneous seizures can arise more easily.

While the mechanisms underlying many of the other known causes for spontaneous seizures are less well characterized, it is thought that all must involve an increase in excitability of neurons, either through the GABAA system or through direct increase in excitability through excitatory neurotransmitter receptors like the NMDA receptor. While sound in theory, there is little direct clinical evidence that NMDA receptor blockade leads to improved seizure control or reductions in seizure risk. However, autoantibodies that bind to the NMDA receptor have been implicated in some cases of refractory status epilepticus, providing some data for an important role for this receptor in seizure development and maintenance.


Several studies have tried to identify the clinical features and primary risk factors for status epilepticus. In general, prior brain injury is the single greatest risk factor for the development of seizures. While all forms of brain injury, including trauma, ischemic stroke and intracranial hemorrhage, can lead to seizures, penetrating brain trauma appears to carry the highest risk. The situation is a bit different for patients presenting in status epilepticus. Consistently, noncompliance with antiseizure medication by an epileptic patient is the number-one cause for status patients presenting to the emergency department.

More recently it has been reported that in cases of refractory status epilepticus, new-onset status is more common than recurrent status. This likely reflects the more varied and complex underlying causes associated with refractory status cases. In general, highly refractory seizures are commonly seen in cardiac arrest patients, patients with subdural hemorrhages, and in prion diseases. Less commonly seen causes for refractory cases are autoimmune diseases and paraneoplastic syndromes. Prior brain injury is considered a risk factor and can also lead to status.


As one might expect, the prognosis is most tightly correlated to the underlying cause for the seizures. In the ICU setting, most patients with ongoing seizures have a clear underlying cause, either infectious or structural. In some, imaging is normal and the seizures are new in onset and the overall recovery potential of the patient is tightly linked to the time it takes to stop the seizures and what is required to keep the patient from going back into status.

There is clear clinical evidence that the longer generalized seizures continue, the harder they are to stop and the higher the mortality. This emphasizes the need for early screening and aggressive intervention. While early detection of seizures in the ICU using continuous EEG has not been shown to improve outcomes, it has been found to identify subsets of patients who have higher morbidity and mortality. In the future, these patient subgroups may become the focus of new interventions and clinical research efforts aimed at finding treatments that can improve these outcomes.

What's the evidence?

Description of the problem

Young, GB. “Continuous EEG monitoring in the ICU”. Acta Neurol Scand. vol. 114. 2006 Jul. pp. 67-8. This is a great review article providing an overview of the evidence that continuous EEG monitoring is useful in the ICU population.

Hirsch, LJ. “Continuous EEG monitoring in the intensive care unit: an overview”. J Clin Neurophysiol. vol. 21. 2004 Sep-Oct. pp. 332-40. This article also provides a nice discussion of several of the early reports of non-convulsive status in hospitalized patients.

Jordan, KG. “Continuous EEG and evoked potential monitoring in the neuroscience intensive care unit”. J Clin Neurophysiol. vol. 10. 1993 Oct. pp. 445-75. This article looked at how EEG data affect care decisions in the ICU, even in cases where seizures are not found.

Emergency Management

Simple concise reviews of the evidence for the current management approach to status epilepticus:

Rüegg, SJ, Dichter, MA. “Diagnosis and Treatment of Nonconvulsive Status Epilepticus in an Intensive Care Unit Setting”. Curr Treat Options Neurol. vol. 5. 2003 Mar. pp. 93-110.

Varelas, PN. “How I treat status epilepticus in the Neuro-ICU”. Neurocrit Care.. vol. 9. 2008. pp. 153-7.

Minicucci, F, Bellini, A, Fanelli, G, Cursi, M, Paleari, C, Dylgjeri, S, Comi, G. “Status epilepticus”. Neurol Sci. vol. 27. 2006 Mar. pp. S52-4.


These articles simply demonstrate that EEG is required to make the diagnosis of seizures in the ICU patient

Oddo, M, Carrera, E, Claassen, J, Mayer, SA, Hirsch, LJ. “Continuous electroencephalography in the medical intensive care unit”. Crit Care Med. vol. 37. 2009 Jun. pp. 2051-6.

Young, GB. “Continuous EEG monitoring in the ICU: challenges and opportunities”. Can J Neurol Sci. vol. 36. 2009 Aug. pp. S89-91.

Drislane, FW, Lopez, MR, Blum, AS, Schomer, DL. “Detection and treatment of refractory status epilepticus in the intensive care unit”. J Clin Neurophysiol. vol. 25. 2008 Aug. pp. 181-6.

Bearden. “Newer data on the frequency of seizures in comatose patients in the ICU with no overt evidence of seizures”. Am. J. END Technol.

Alroughani, R, Javidan, M, Qasem, A, Alotaibi, N. “Non-convulsive status epilepticus; the rate of occurrence in a general hospital”. Seizure. vol. 18. 2009 Jan. pp. 38-42. This article reinforces the fact that clinical symptoms cannot be used to exclude the possibility of underlying seizures:

Specific Treatment

These articles simply provide a few interesting reports on autoimmune causes of SE and their clinical course and treatment:

Kirkpatrick, MP, Clarke, CD, Sonmezturk, HH, Abou-Khalil, B. “Rhythmic delta activity represents a form of nonconvulsive status epilepticus in anti-NMDA receptor antibody encephalitis”. Epilepsy Behav. 2010 Dec 27.

Milh, M, Villeneuve, N, Chapon, F, Gavaret, M, Girard, N, Mancini, J, Chabrol, B, Boucraut, J. “New onset refractory convulsive status epilepticus associated with serum neuropil auto-antibodies in a school aged child”. Brain Dev. 2010 Nov 12.

Johnson, N, Henry, C, Fessler, AJ, Dalmau, J. “Anti-NMDA receptor encephalitis causing prolonged nonconvulsive status epilepticus”. Neurology. vol. 75. 2010 Oct 19. pp. 1480-2.

Dalmau, J. “Status epilepticus due to paraneoplastic and nonparaneoplastic encephalidides”. Epilepsia. vol. 50. 2009 Dec. pp. 58-60.

Disease monitoring, follow-up and disposition

These articles provide data on outcomes from convulsive and non-convulsive status epilepticus and reference other primary study literature:

Legriel, S, Mourvillier, B, Bele, N, Amaro, J, Fouet, P, Manet, P, Hilpert, F. “Outcomes in 140 critically ill patients with status epilepticus”. Intensive Care Med. vol. 34. 2008 Mar. pp. 476-80.

DeLorenzo, RJ, Garnett, LK, Towne, AR, Waterhouse, EJ, Boggs, JG, Morton, L, Choudhry, MA, Barnes, T, Ko, D. “Comparison of status epilepticus with prolonged seizure episodes lasting from 10 to 29 minutes”. Epilepsia. vol. 40. 1999 Feb. pp. 164-9.

Towne, AR, Waterhouse, EJ, Boggs, JG, Garnett, LK, Brown, AJ, Smith, JR, DeLorenzo, RJ. “Prevalence of nonconvulsive status epilepticus in comatose patients”. Neurology. vol. 54. 2000 Jan 25. pp. 340-5.