Status Epilepticus

Genetic epilepsy is one in which a known or presumed genetic defect(s) results in seizures which are the core symptom of the disorder. A structural/metabolic epilepsy is one in which there is a demonstrated increased risk of epilepsy with the structural or metabolic condition.

Typical and atypical absence seizures: Absence and atypical absence may actually be several different types of SE, which have a similar presentation. Both absence and atypical absence may be seen in genetic epilepsies, and are terminated by AEDs. In generalized structural/metabolic epilepsies, there may be overlap with focal SE, due to frontal focal lesions. In the elderly, new-onset absence SE may be seen; there are also drug-induced, and drug-withdrawal versions of absence SE. The ILAE terminology does not capture some significant details of these events, which make them appear to be different phenomena. However, there are five well-described versions of absence seizures, given below.

Absence status epilepticus is typically considered to be a component of genetic epilepsy, with impairment of consciousness. The level of impairment is often “individual-variable,” with about 20% having slight clouding of consciousness, about 60% a confusional state in which the patient is typically calm but does not interact with the environment, and about 20% with more severe impairment; there are sometimes accompanying subtle jerks of the eyelids during the event. The EEG correlate is bilateral and symmetric, typically bifrontally predominant, spike or polyspike and wave complexes at least 2.5 Hz, at least initially during the event; other patterns are also possible. This type of status recurs in most patients, and rarely can occur frequently (16). Neuronal damage is unlikely to occur with this type of status, and as such, aggressive treatment is not recommended (17). Most commonly, intravenous benzodiazepines will terminate the event. If ineffective, one may consider using intravenous valproic acid (18). Additionally, valproic acid may be effective in reducing recurrent events in patients with multiple episodes of absence status.
Atypical absence status is more commonly encountered in patients with structural/metabolic epilepsy, presenting with a fluctuating level of consciousness; it is also seen in patients with genetic epilepsy. This fluctuating confusional state is different than absence SE, which usually has a certain level of impairment. The ictal semiology is quite different than absence status, because it can include tonic, atonic, myoclonic, or otherwise lateralized phenomena. The EEG is spike, and polyspike and wave complexes, which are irregular, but even when quasirhythmic, occur at less than 2.5 Hz; these episodes may recur. The atypical episodes are generally not amenable to benzodiazepine treatment. In patients with recurrent atypical absence SE with an underlying genetic epilepsy, valproic acid may be particularly helpful in reducing recurrences. Atypical absence SE is likely a form of status not causing neuronal damage (19).
Absence SE with focal features is most typically encountered in frontal lobe localization-related epilepsy. There is impairment of consciousness, but the level of impairment may be individual dependent. The EEG is typically bilateral, but asymmetric and may develop into one looking like absence SE later in the episode. Treatment response varies with individual.
Late-onset, de novo absence SE occurs in older adults, with an underlying toxic or metabolic issue leading to seizures. Such patients can have repeated episodes with recurrent toxic/metabolic issues causing further episodes of SE. The preferred treatment is to deal with—including prevention—the underlying toxic/metabolic cause, but the individual episode can be typically easily terminated with benzodiazepines (18). Of note, there may not be a need to treat these patients with long-term AED therapy (20).
Myoclonic absence seizures are proximal, predominantly upper extremity myoclonic jerks synchronized to the 3 Hz spike and wave seen on the EEG during the SE. These can last for hours or days, and are most commonly refractory to therapy. Treatment is possible for some patients, and, in particular, withdrawal of agents known to aggravate idiopathic generalized epilepsies (like carbamazepine) may be beneficial (21).

Myoclonic seizures do not cause a change in consciousness. There is irregular, typically bilateral, myoclonic jerking which may persist for hours. Seen in conjunction with Dravet syndrome, myoclonic astatic epilepsy, nonprogressive myoclonic epilepsy in infancy (especially Angelman syndrome), and incompletely controlled juvenile myoclonic epilepsy. It can be benign, without untoward sequelae (22).
The ILAE report makes no mention as to whether negative myoclonic status, like that which may be seen in continuous spike-wave in slow wave sleep qualifies as myoclonic seizures; there is a logic to including it in this part of the classification schema. In myoclonic SE, there is a limb, often in the upper extremity, which becomes paralyzed, but has continued, brief atonic episodes. There can be alteration of consciousness with these events, which can appear during the episode, with risk that the abnormalities may persist after the episode ends. Myoclonic seizures are associated with anoxic encephalopathy. In the era before induced hypothermia was used, the appearance of these seizures heralded a dire prognosis; this may no longer be the case.

In tonic SE, the patient will have brief tonic spasms, which can continue for hours, interspersed with periods of apparent calm. Most typically, if the patient is lying down, the neck and arms are flexed. Tonic SE may occur with both structural/metabolic and genetic epilepsies; with structural/metabolic seizures, the duration can be longer than hours.

Subtle SE is the end result of uncontrolled tonic–clonic SE, with focal or multifocal myoclonias, coma, periodic lateralized epileptiform discharges (PLEDs), and a slow suppressed background EEG (23,24); the myoclonias may not be epileptic in nature. The ILAE guidelines do not provide details of the myoclonias, but they are typically in the form of subtle twitches of the trunk or extremities, and can also present as nystagmus. The EEG typically is composed of ictal, but asymmetrically bilateral, rhythmic discharges. While the ILAE guidelines are silent about progression, eventually, there is complete loss of motor component, with only ongoing ictal EEG activity; electrocerebral silence ensues if the seizure is not controlled.

Nonconvulsive status epilepticus (NCSE) is not part of the ILAE definition, yet is likely more frequently encountered in properly monitored ICU patients than all other forms of SE. While NCSE rarely appears as the end stage of tonic–clonic SE, it is far more frequently found in ICU patients with unexplained mental status. While NCSE is an area of ongoing investigation, a frequently quoted definition is given as follows (25).

In patients without a known epileptic encephalopathy:
1. Rhythmic spikes, polyspikes, sharp waves, sharp and slow wave complexes of greater than 2.5 Hz
2. Rhythmic spikes, polyspikes, sharp waves, sharp and slow wave complexes of less than 2.5 Hz, or rhythmic delta/theta activity of greater than 0.5 Hz with one or more of the following.
  1. Intravenous AEDs result in improvement of both clinical status and EEG.
  2. During the EEG pattern above, there are subtle clinical ictal phenomena.
  3. There is an increase in voltage and change in frequency at onset, with a change in frequency by more than 1 Hz and/or change in location/spread during event, or a change in voltage or frequency on termination of the event.
In patients with known epileptic encephalopathy:
1. There is an increase in prominence or frequency in the features mentioned in 10A (above), with both an observable change in clinical status and a change from baseline at the same time.
2. Intravenous AEDs cause improvement of both clinical status and EEG status. It is an open question as to how aggressively NCSE should be treated.

Febrile SE was included in the 1993 ILAE report, but is not mentioned in the 2006 report, in which the ILAE explicitly classifies SE. We include it herein for purposes of completeness. Febrile SE is defined as seizure during a febrile illness without a central nervous system (CNS) infection or acute electrolyte imbalance in a child older than 1 month of age, without previous febrile seizures, in which there are 30 minutes of continuous seizure activity, or intermittent activity lasting at least 30 minutes without return to baseline consciousness. There is evidence that outcome is related to the duration of the seizure; in particular, seizure duration longer than 120 minutes is associated with poor outcome (26).

EPIDEMIOLOGY

The incidence of SE ranges from 10 to 41 per 100,000 individuals per year (27–30). All studies showed a significantly higher incidence in the elderly, especially after 60 years of age, raising a concern that the overall incidence may rise as the population ages. In addition, only 40% to 50% of patients presenting with first-time SE have a previous diagnosis of epilepsy (28,29). NCSE represents about 30% to 40% of all cases of SE, with an estimated incidence of 5 to 9 per 100,000 individuals per year. However, the true incidence of NCSE may be underestimated. In fact, in various studies the reported incidence of NCSE in ICU patients with altered mental status ranged from 8% to 37% (31–35). Diagnosis requires clinical suspicion and long-term EEG monitoring, which is not routinely performed on critically ill patients in many institutions.

Mortality from SE, estimated in most studies at 10% to 20%, rises significantly with age (36), reaching 38% in the elderly (60 years of age or greater) (27). One of the primary predictors of poor outcome is prolonged seizure; a seizure lasting more than 1 hour has mortality reaching 32%, compared to 2.7% with shorter seizures (36,37). Mortality from NCSE seems to be higher, averaging 50% (38).

ETIOLOGY

In about 30% of cases, SE occurs in patients with chronic epilepsy and is due to withdrawal, or low blood concentrations, of AEDs (37,39,40). In the majority of cases, SE occurs in patients with no history of epilepsy and may be due to a variety of causes, most commonly intracranial pathology, such as ischemic stroke, intracerebral (ICH) and subarachnoid hemorrhage (SAH), CNS infections, head trauma, and brain tumors. Other etiologies include cardiac arrest and hypoxic/anoxic brain injury, alcohol-withdrawal, metabolic disturbances, and toxic causes. In some patients, no cause is identified (39,40).

Both acute and chronic intracranial pathology can cause seizures. Seizures and SE may actually be the presenting signs of several neurologic conditions. This is true for intracranial hemorrhage, including SAH and ICH, acute embolic stroke, and brain tumors. Approximately 50% of patients with brain tumors experience seizures (41,42), and a seizure is the presenting sign of a tumor in 23% of cases (43). Seizures can also be the presenting sign of an acute stroke (44) and frequently occur in the first 2 weeks after a stroke. It is estimated that seizures occur in up to 6% of patients with ischemic stroke, up to 18% of patients with ICH, and up to 26% of patients with SAH (44,45). Up to 2.8% of patients with stroke go into SE either at presentation, or within 2 weeks, of their stroke (46). The risk of chronic epilepsy is 17 times higher after an ischemic stroke than the general population (47), and the risk of having a seizure or developing chronic epilepsy after any type of stroke is 11.5% (48). In SAH, generalized tonic–clonic seizures have been reported in up to 26% of patients at the time of onset or shortly after onset (45,49), and NCSE occurred in 8% of patients who survived the first 48 hours and had an unexplained decline in their level of consciousness (50).

Metabolic disturbances that may cause seizures include hyponatremia, hypoglycemia, hypocalcemia, hypomagnesemia, uremia, hepatic encephalopathy, and hyperosmolar states (50). However, it is important to note that metabolic encephalopathies can frequently cause EEG abnormalities that can be difficult to distinguish from subtle seizure activity, such as high-amplitude slowing and triphasic waves. Therefore, extra care should be taken to avoid both over- and underdiagnosing patients as having SE when they have a clear metabolic dysfunction; response to treatment may be critical in these situations.

Several drugs can cause seizures at toxic levels, including some analgesics such as meperidine, propoxyphene, and tramadol; some psychiatric medications such as bupropion, tricyclic antidepressants, lithium, olanzapine, selective serotonin reuptake inhibitors (SSRIs), venlafaxine, and clozapine. Theophylline, isoniazid, lidocaine, phenothiazines, and some antibiotics such as imipenem/cilastatin, penicillins, and ciprofloxacin may also induce seizures. Furthermore, several commonly abused drugs can cause seizures, most notably cocaine, amphetamines, phencyclidine, and γ-hydroxybutyric acid (51,52).

PATHOPHYSIOLOGY AND MECHANISMS

The great majority of seizures stop spontaneously in less than 2 minutes (53). This is most likely due to inhibitory mechanisms that attempt to deter any excessive, abnormal neuronal activity. This inhibition is evident on the EEG as postictal slowing and attenuation. It is believed that SE occurs when inhibitory mechanisms fail, resulting in a self-sustaining and prolonged seizure activity; the exact cause of this failure is not well understood. A large number of elegant experiments done on animal models of SE have attempted to shed light on the underlying mechanisms causing SE. Review of these studies is beyond the scope of this chapter; however, two points are worth discussing, since they have important implications on treatment strategy.

Self-sustaining SE can be easily triggered in animal models using electrical stimulation (54). However, this can be blocked by many drugs that increase inhibition or reduce excitation only if administered early, prior to the development of a self-sustained seizure (55). In contrast, once a self-sustaining state is established, it becomes more difficult to stop the seizure (56), and much higher dosages of inhibitory drugs are required, leading to significant toxicity, including cardiovascular depression (57). Another important feature of self-sustaining SE is the progressive development of resistance to AEDs. The anticonvulsant potency of benzodiazepines can decrease by 20 times within 30 minutes of self-sustaining SE (58). The same phenomenon was observed with other anticonvulsants, such as phenytoin; however, the decline in potency was slower (59).

Pathophysiologically, SE produces a number of neurologic and systemic changes. Primary neurologic complications occur in both convulsive and some forms of nonconvulsive SE, and are time dependent and probably preventable with early termination of the seizure. In animal models of SE, neuronal injury occurs even in the absence of convulsive activity (60,61), and cell death is thought to result from excessive neuronal firing through excitotoxic mechanisms (62). It is impossible to replicate these experiments in human beings; however, there is widespread belief—supported by some anecdotal evidence—that neuronal injury and death occur after prolonged seizures. For example, brain damage and decreased hippocampal neuronal density are often seen in patients who die from SE (63,64). Furthermore, cerebral edema and chronic brain atrophy seen on neuroimaging studies have been reported after SE (65–68).

Systemic complications of prolonged seizures are seen primarily in GCSE, and are due to autonomic hyperactivity and excessive muscle activity. Therefore, systemic complications can potentially be prevented, or minimized, with early termination of seizure activity or induction of muscle paralysis and artificial ventilation (61). Pathophysiologic manifestations include increased systemic blood pressure, tachycardia, and cardiac arrhythmias; increased pulmonary blood pressure; increase in cerebral blood flow; elevation of body temperature; increased peripheral white cell count; transient pleocytosis in the spinal fluid; and a marked metabolic acidosis (60,69,70). Epinephrine levels are elevated and reach the dysrhythmogenic range; these may play a role in sudden death (70). With prolonged convulsive SE—defined as lasting 30 minutes or more–systemic blood pressure and cerebral blood flow can drop significantly (60). Additionally, blood glucose is initially elevated in response to excessive adrenergic stimulation; however, after 30 minutes of GCSE, hypoglycemia may occur (60). Both hypoglycemia and decreased cerebral blood flow contribute to further neuronal injury (71). Excessive muscle contraction often causes severe metabolic acidosis, breakdown of muscle tissue, and hyperkalemia (60,61,69). Arterial pH has been reported to fall below 7.0 (72) and contribute, along with hyperkalemia, to cardiac dysrhythmias. Rhabdomyolysis and myoglobinuria can also occur and may lead to acute renal failure (73).

EVALUATION

Clinical Presentation

Obtaining a focused history and examination may be very helpful for diagnosis and management (Table 121.1). Convulsive and nonconvulsive SE have very different clinical presentations, and their treatment is quite different. Convulsive SE frequently occurs outside the hospital, and management may start in the ambulance before patients arrive to the emergency room. The diagnosis is usually evident, unless there is a strong clinical suspicion of psychogenic nonepileptic seizure (PNES). Convulsive SE often starts as a focal seizure with secondary generalization. Rarely, primary generalized seizures evolve into SE. The generalized convulsion either becomes continuous, or stops and recurs before the patient regains full consciousness. In either case, the tonic–clonic activity changes in character with time and often patients go into a continuous clonic phase where clonic activity persists and gradually slows down and becomes more subtle. With time, the only persistent motor activity may consist of small-amplitude twitching of the face, hands, or feet or nystagmoid jerking of the eyes (74,75). Sometimes the motor activity subsides completely, and patients remain stuporous or comatose; in this case, patients evolve from convulsive to NCSE (33).

TABLE 121.1 History and Physical Examination

By the time patients arrive to the emergency department (ED), they may already be in established SE. If there is strong clinical suspicion of PNES, an EEG is essential to confirm the diagnosis. The average duration of a PNES in one study was 5 minutes (76); however, they can be protracted and may mimic convulsive SE. Another study of patients in an epilepsy monitoring unit had nearly 20% of PNES patients, with PNES mimicking SE (77). Additionally, patients with PNES may also have epilepsy, making it all the more difficult to know how to treat.

NCSE has a different clinical presentation, with unexplained decline in mental status that cannot be completely explained by other causes. It may occur either outside the hospital or, frequently, in the hospital, in patients already admitted for other reasons such as stroke, intracranial hemorrhage, brain tumors, or metabolic disturbances. Frequently, the underlying etiology may account in part for the impairment in consciousness; however, patients frequently have an unexplained decline of mental status after a period of clinical improvement. Therefore, clinical suspicion should be strong, and evaluation for NCSE should be undertaken in any patient with unexplained impairment in mental status.

Electroencephalogram

The EEG is the only diagnostic tool that can confirm or refute the diagnosis of SE. In GCSE, an EEG may not be necessary initially, unless PNES must be excluded. However, if convulsive activity stops and patients do not recover their baseline level of consciousness, evaluation with an EEG is important to exclude the continuous presence of seizure activity. In NCSE, the EEG is essential. However, a single routine EEG of 20 minutes’ duration may not be adequate and may only capture seizure activity in 20% of cases. A longer EEG recording of at least 1 hour increases the sensitivity to 50%. More prolonged EEG monitoring is recommended if shorter-duration EEGs are nondiagnostic. Long-term EEG monitoring of 24 to 48 hours can increase the diagnostic accuracy to over 90% (35). Several EEG patterns have been described during SE, probably reflecting different stages of brain activity (75). In addition, several patterns have been described in NCSE. Discussion of these different EEG patterns is beyond the scope of this chapter; however, an important issue needs to be emphasized. Some EEG patterns can be difficult to distinguish from epileptiform activity, such as diffuse triphasic waves in metabolic encephalopathies (Fig. 121.1) and breach rhythms after a craniotomy (Fig. 121.2). These patterns can be very deceiving and can often be misinterpreted as epileptiform. Therefore, it is very important for the EEG to be interpreted by an experienced electroencephalographer.

Neuroimaging

Neuroimaging studies are always recommended to assess for the presence of intracranial pathology. Even in patients with known pathologies, such as tumors or stroke, repeat imaging is recommended to exclude progression or complications of the underlying disease. For example, a stable tumor can become necrotic or hemorrhagic, or a stable acute or subacute infarct can turn hemorrhagic. Unenhanced computed tomography (CT) of the brain is adequate in the acute setting; however, magnetic resonance imaging (MRI) is much more sensitive and may detect lesions not seen on CT.

Laboratory Evaluation

Full laboratory evaluation is always recommended (Table 121.2), including blood cell count, renal function, liver function, electrolytes, calcium, magnesium, and AED levels. Toxicology should be performed when there is a clinical suspicion of intoxication or substance abuse. This is especially important in patients with a psychiatric illness at risk of suicide and in children who may have access to adult medications. Lumbar puncture is indicated if there is any consideration of an infectious etiology. Also, a lumbar puncture should be considered when SAH, not seen on CT scan, is suspected. However, in the presence of any sign of intracranial hypertension, lumbar puncture should be avoided, since it may increase the risk of transtentorial herniation. It is important to note that patients with convulsive SE often exhibit clinical features suggestive of meningitis, such as elevated temperatures, increased peripheral white blood cell counts, and pleocytosis in the cerebrospinal fluid (CSF) (70). These abnormalities have been reported in up to 18% of patients with convulsive SE, without any evidence of infection (70), and are thought to result from breakdown of the blood–brain barrier. Usually, the total white blood cell count in the CSF remains under 100 and glucose level remains normal. Treatment with antimicrobials should be initiated if there is clinical suspicion for a CNS infection.

TREATMENT

General Rubric

The most important thing to know is that this is a very complicated subject. Each of the 11 different kinds of status is considered differently. It can be confusing, when looking at papers on this subject, because there is often the assumption that when an author talks about SE, he or she means generalized convulsive SE.

Generalized Convulsive Status Epilepticus

Treatment Principles

GCSE is a medical emergency and should be dealt with as such. Therapies are aimed at early termination of seizure activity, identification and correction of the cause, prevention of seizure recurrence, and treatment of pathophysiologic complications. There is ample evidence that delayed treatment leads to poor outcome (36,78). In addition, there is a time-dependent loss of efficacy of anticonvulsant medications (58,59). Therefore, early initiation of aggressive treatment is essential in the management of GCSE. It is highly recommended that every ED and ICU have a well-defined and clear treatment protocol. This helps avoid many of the pitfalls leading to delayed and insufficient treatment of SE (79).

Prehospital Management

In many cases, patients with convulsive SE are brought into the ED by ambulance, making prehospital treatment possible. Initiation of treatment in the ambulance is highly recommended, when possible, given the importance of early intervention. Both rectal diazepam (80,81) and intravenous diazepam and lorazepam (82) can be safely and effectively used. In one randomized, double-blind, prospective study (82), seizures terminated before arrival to the ED in 59% of patients who received intravenous lorazepam, 43% of those who received intravenous diazepam, and 21% of those who received placebo. The safety profile was also good, with more patients having respiratory or circulatory complications in the placebo group than the treatment groups. Treatment in the ambulance with intravenous benzodiazepines should only be initiated if the paramedical team transporting the patient has the training and equipment to perform endotracheal intubation and artificial ventilation, in case of respiratory depression.