How can the physician distinguish between simple partial seizures and complex partial seizures?

In most cases, the two basic types of partial seizures are easy to differentiate.

Simple partial seizures

The signs or symptoms of simple partial seizures depend on the location of the seizure focus. Seizures involving the motor cortex usually are not difficult to diagnose. They commonly consist of rhythmic or semi-rhythmic clonic activity of the face, arm, or leg.

Seizures with somatosensory, autonomic, and psychic symptoms (hallucinations, illusions, déja vu) may be more difficult to diagnose. Psychic symptoms more often are a component of a complex partial seizure.

Complex partial seizures

Complex partial seizures (CPS) are one of the most common seizure types encountered in both children and adults.

By definition, all patients with CPS have impaired consciousness. The patient either does not respond to commands or responds in an abnormally slow manner. Simple staring and impaired responsiveness alone may characterize a CPS, but usually behavior during the seizure is more complex. Automatisms (involuntary motor activity) are common during the period of impaired consciousness. This behavior is quite variable. Common activities include:

• facial grimacing
• gesturing
• chewing
• lip smacking
• snapping fingers
• repeating phrases

The patient does not recall this activity after the seizure.

The length of CPS varies. They usually last from 30 seconds to several minutes. Their length helps to differentiate them from absence seizures, which usually last less than 15 seconds.

Most patients have some degree of postictal impairment, such as tiredness or confusion, after CPS.

Complex partial seizures may be preceded by a simple partial seizure (aura), which can serve as a warning to the patient of a more severe seizure to come. The nature of the aura may enable the physician to determine the cortical area in which the seizure is beginning.

What are the types of generalized seizures?

Primary generalized seizures arise simultaneously in both cerebral hemispheres. They generally involve impairment of consciousness, but their manifestations can range from barely noticeable to a tonic-clonic convulsion.

Sometimes it is difficult to distinguish primary generalized seizures from secondarily generalized seizures that are preceded by a very brief simple partial seizure, especially since the patient may have no memory of it.

Types of generalized seizures include:

Myoclonic seizures are characterized by sudden, brief (<350 milliseconds), shock-like contractions.

The contractions may be:

• generalized
• confined to the face and trunk
• confined to one or more extremities
• confined to individual muscles or groups of muscles.

Any group of muscles can be involved in the jerk. Myoclonic seizures may be dramatic, causing the patient to fall to the ground, or be quite subtle, resembling tremors. Because of the brevity of the seizures, it is not possible to determine if consciousness is impaired.

Generalized tonic-clonic seizures

Generalized tonic-clonic (GTC) seizures (formerly called grand mal seizures) usually are not difficult to diagnose. The loss of consciousness usually occurs simultaneously with the onset of a generalized stiffening of flexor or extensor muscles (the tonic phase). After the tonic phase, generalized jerking of the muscles (clonic activity) occurs. A GTC seizure is almost always followed by deep postictal sleep.

Some patients may have a simple partial seizure (aura) preceding the loss of consciousness. This indicates that the onset of the seizure is located in a limited area. As the seizure spreads in the cortex, the seizure develops into a GTC seizure. This type of seizure is classified as a simple partial seizure with secondary generalization.

Two types of episodes that may end with brief generalized tonic-clonic seizures should not be treated with antiepileptic drugs:

• breath-holding attacks in toddlers
• syncope in both children and adults

Atonic (astatic) seizures, or drop attacks, are characterized by a sudden loss of muscle tone. They begin suddenly and without warning. There may be total lack of tone, so that the person will fall quickly to the floor and cannot protect himself or herself against injury. Consciousness is impaired during the fall, although the patient may regain alertness immediately upon striking the floor.

Some atonic seizures may be fragmentary and lead to dropping of the head with slackening of the jaw or dropping of a limb.

Electromyographic activity is lost during atonic seizures.

What are nonepileptic seizures?

Nonepileptic seizures (NESs) are episodic paroxysmal events that resemble epileptic seizures in many respects. Such events are often difficult, if not impossible, to differentiate from events due to epilepsy, and misdiagnosis leads to inappropriate treatment with antiepileptic drugs (AEDs).

To arrive at an accurate diagnosis, the clinician first must differentiate between the two major types of NESs:

• Psychogenic NESs are symptoms of an underlying psychiatric disorder, without a physiologic basis.
• Physiologic NESs are caused by physiologic dysfunction, such as cardiac arrhythmias, hypotensive episodes, or cerebrovascular disease. Such conditions may result in loss of consciousness, with or without associated motor manifestations. A detailed history and appropriate investigations (e.g., Holter monitoring, noninvasive carotid artery studies, or tilt-table testing) will usually reveal the true diagnosis.

Most of the discussion in this part of epilepsy.com pertains to psychogenic NESs.

Why is an understanding of nonepileptic seizures important?

NESs are relatively common. As many as 20% of the population seen in specialized seizure clinics may suffer from NESs. Put another way, about 50,000 persons in the United States have NESs.

NESs may remain undiagnosed for years. If the patient has a past history of epilepsy and experiences a recurrence of seizures or currently has epileptic seizures and a new seizure type develops, the true diagnosis may be obscured. The clinician usually does not witness the seizures and must rely on information provided by the patient, an outside observer, or both.

Mistakenly attributing NESs to epilepsy generally leads to futile treatment with AEDs. Continuing seizures in spite of increasing AED dosage or multiple AEDs is frustrating for patient and physician alike. Moreover, AED therapy may lead to toxic side effects, causing additional disability and frustration. A correct diagnosis will lead to appropriate psychiatric intervention, discontinuation of AED therapy, and improvement in the patient's quality of life.

What are the most common psychiatric diagnoses associated with nonepileptic seizures?

Nonepileptic seizures are classified as a conversion disorder in the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV). Conversion disorder is included in the broader category of somatoform disorders. Essentially, the patient presents with symptoms suggestive of a neurologic or other general medical condition, preceded by conflicts or other stressors. The symptom is not feigned, and appropriate investigations fail to reveal evidence of a causative organic condition. The symptoms cause significant distress and interfere with the patient's general functioning.

In some patients, NESs are part of a symptom complex subsumed under the rubric somatization disorder. In this condition, the patient has a pattern of recurring multiple and significant somatic complaints beginning under the age of 30 and extending for a prolonged period, sometimes many years.

Patients with NESs also suffer from associated psychiatric disorders. More than one condition is often present. Anxiety disorders are commonly encountered, often unrecognized by patient or physician. In particular, many patients fulfill DSM-IV diagnostic criteria for panic disorder, with or without agoraphobia. The symptom complex of NESs may be due solely to panic attacks, or the attacks may coexist with NESs. Careful inquiry is essential in order to establish the diagnosis.

Depression is often seen in patients with NESs. In some, it has been proposed that the pain of depression has been unrecognized or unaddressed by family or others. As a result, the development of NESs may constitute a mechanism of bringing the patient's problems to the attention of the medical profession.

Psychotic disorders such as schizophrenia are considered uncommon accompaniments of NESs, although they may be seen.

Other conditions that are even less common include malingering and factitious disorder.

A detailed psychiatric interview will usually bring out any associated psychiatric disorders. In some cases, however, application of the Structured Clinical Interview for DSM-IV (SCID) may be required to determine the number and extent of the patient's problems.

Numerous studies have evaluated the clinical differentiation between nonepileptic seizures (NESs) and epileptic seizures. This section addresses the evaluation, diagnosis and treatment of NESs, and provides a guide for discussing the diagnosis of NES with patients.

Generally, there are many differences between the histories of patients with nonepileptic seizures (NESs) and those of patients with epilepsy.

NESs may occur only in the presence of others or, conversely, may never have been observed. In the latter case the clinician is dependent on the patient's description, which may be fragmentary and incomplete. Indeed, only loss of consciousness may be reported. If such a patient has a history of epilepsy, the probability of misdiagnosis is high.

The interview may reveal a clear emotional trigger for NESs. This criterion is not at all definitive, however, because patients with epilepsy often ascribe seizures to "stress."

The patient's reaction to his or her seizures may offer clues to the diagnosis. The demeanor of the patient with NESs classically has been described as la belle indifference, but many authors have found such indifference to be atypical. More commonly, the patient is quite concerned about the seizures, sometimes excessively so. In fact, an exaggerated emotional response may provide a clue to NESs but should be considered in context with other information.

Treatment with AEDs, regardless of types or combinations, is rarely successful, although transient responses are not uncommon. In fact, increasing doses of AEDs may lead to a paradoxical increase in seizure frequency. Patients with NESs sometimes complain of intolerable side effects at low doses or slow dose escalation of AEDs.

A history of physical or sexual abuse is common in patients with NESs. At an appropriate time the clinician should inquire into this sensitive subject. Drawing firm conclusions from a history of abuse is perilous, however, because abuse is common in people with epilepsy and in the general population. Many people with epilepsy also have poignant histories of childhood or marital abuse.

Finally, suspicion is sometimes kindled by the patient's previous experience with other people who have epilepsy. The person may have encountered seizures in a professional capacity-for example, in a hospital setting. A family member or friend may have seizures, or the patient's own previous or current epileptic seizures may serve as templates for NESs. In addition, cultural influences may play a role in some cases.

Do patients with nonepileptic seizures have a typical psychologic profile?

Patients with NESs have no characteristic psychologic profile. Perhaps the most consistent results have resulted from the application of the Minnesota Multiphasic Personality Inventory (MMPI). In particular, Dodrill has reported that MMPI profiles differ between patients with epilepsy and those with NESs (Dodrill 1993). The typical findings in many NES patients are relatively high scores on the hysteria and hypochondriasis scales, with a lower score on the depression scale. These characteristics differentiated NESs from epilepsy in about 80% of cases.

Thus, the MMPI offers useful information but cannot be said to have sufficient power for diagnostic certainty. The results of psychologic testing, therefore, must be taken in concert with the results of other testing and considered supportive or nonsupportive of the diagnosis of NES.

If the clinician observes one or more NESs, the phenomenology of the event may raise suspicion that the patient does not have epilepsy. Nothing in the clinical expression of NESs can be considered definitively pathognomonic, however. Only by recording the behavioral and electrographic characteristics can the event be diagnosed confidently.

One common clue is that NESs often have a gradual onset, but epileptic seizures typically start and end suddenly. For example, a simple absence seizure (formerly known as petit mal) begins with sudden loss of awareness, which continues for several seconds and then ends abruptly, at which time the person is fully aware and functional. Similarly, a tonic-clonic epileptic seizure begins suddenly with a fall, loss of consciousness, and frequently a cry. Motor activity continues for a minute or so, progressing predictably and ending abruptly. At that point the patient is deeply comatose and flaccid. With NESs, on the other hand, consciousness may be preserved during the peri-ictal phase.

In contrast, a nonepileptic seizure characterized by vigorous motor activity frequently builds in intensity as the event progresses. The event itself is often of longer duration than its epileptic counterpart, sometimes lasting many minutes or even hours. Moreover, the motor movements may have a waxing and waning quality through the course of the seizure. (Some NESs with motor manifestations are characterized by vigorous motor activity and intervening periods of quiet unresponsiveness.) The event tends to subside gradually, and the postictal state-whether unresponsiveness or apparent confusion-is less profound than in the case of a tonic-clonic convulsion.

Although it is often said that tongue biting and incontinence do not occur with NESs, some patients with documented NESs have reported these symptoms. Bladder and bowel incontinence and self-injury are rare but can occur. When biting occurs, it tends to involve the tongue tip, arms, or other body areas more than the sides of the tongue.

NESs that resemble true absence attacks are more difficult to differentiate from epilepsy than motor events. The staring, unresponsive state provides no obvious clues concerning the nature of the attack. However, if such attacks are of long duration, one might suspect that the diagnosis is other than epilepsy. The only way to confirm the true diagnosis is to record the event in question by EEG monitoring, preferably with simultaneous video monitoring.

Patients with NESs resembling complex partial seizures also present diagnostic difficulties, owing to the protean manifestations of the epileptic events. Such patients may exhibit confusional states with or without apparent automatic activities. Again, recording of the event is essential to confirm the diagnosis.

Another feature suggestive of NESs is an emotional response during or after an event. Crying and other emotional vocalizations are common. Although crying occurs in rare patients with epilepsy, it is more frequent in those with NES.

Typical NESs, especially motor events, do not seem to follow a "physiologic" progression. The motor activity tends to be chaotic, sometimes with flinging movements. Alternation of the movements, opisthotonus, pelvic thrusting, and dystonic posturing may be present. One important distinguishing feature of NESs is that, in contrast to epileptic seizures, the face is often not involved.

Some patients with NES exhibit avoidance behavior during events, especially during quiescent phases. For example, if the patient's arm is held aloft and released, the patient may avoid striking himself or herself in the face. Such behavior suggests that the attack is nonepileptic.

In some cases NESs can be initiated and ended by suggestion and verbal intervention.

What is the diagnostic role of prolactin levels?

Elevations in serum prolactin occur in the postictal phase (20 to 25 minutes after the seizure subsides) after some types of epileptic seizures. The most consistent increase is found after generalized tonic-clonic convulsions.

The "gold standard" in the diagnosis of nonepileptic seizures (NESs) is a recording of a typical event during video-EEG monitoring. This procedure is available at all centers specializing in epilepsy and is increasingly available at general hospitals and even in some neurologic group practices.

During this procedure, the EEG is recorded for a prolonged period, accompanied by continuous closed-circuit video observation. The digitized EEG and recorded behavior are displayed simultaneously, allowing point-to-point correlations of recorded events and any accompanying electrographic changes.

Types of monitoring

Two types of monitoring are in general use:

• an outpatient procedure with a duration of 6 to 8 hours (Daytime Monitoring, or DAYMON)
• inpatient monitoring, which continues for 24 hours or more

Outpatient studies are less expensive and more convenient than inpatient monitoring. DAYMON is most appropriate for patients with relatively high seizure frequencies-at least three events per week. To increase yield, DAYMON should be carried out when the patient is sleep-deprived.

If the patient's seizure frequency is relatively low, inpatient video-EEG monitoring for 24 hours or more is indicated. This procedure requires hospital admission and a dedicated staff. Although more costly than DAYMON, inpatient monitoring is effective. More than one event may be recorded, increasing diagnostic certainty if the events are stereotyped. Inpatient monitoring also allows recording of a full night's sleep, increasing the possibility of recording sleep-provoked epileptiform activity as well as nocturnal clinical events. Several days of monitoring may be required before the diagnosis is made.

Video-EEG monitoring procedures

During video-EEG monitoring, the patient wears an EEG transmitter connected to a wall outlet by coaxial cable. Wall-mounted video cameras provide continuous behavioral observation. Both EEG and video signals are transmitted to a control room, where the EEG is reformatted and conducted to a video monitor. The EEG signal and video are displayed simultaneously for on-line observation, and both are recorded on videotape. The EEG may be recorded on paper or stored on optical disc.

The patient can move about and carry out normal activities, such as napping, talking, and watching television. Participation by a family member or friend is encouraged, especially someone who has observed the patient's events in the past. Hyperventilation and photic stimulation are carried out. These may cause clinical or diagnostic changes in patients with epileptic seizures but not NESs. Although NESs may occur spontaneously, the application of these procedures appears to increase diagnostic yield.

An important diagnostic aid is suggestion techniques to precipitate one of the patient's usual events. These techniques may take the form of placing alcohol pads over the carotid arteries or administering intravenous saline. The patient is told that the procedure will be carried out to induce a seizure and that only by recording an event will a diagnosis be possible. If an event is precipitated and the event is typical of the patient's usual seizure, a diagnosis of NESs is highly likely. False positives are rare. When DAYMON is performed in this manner, an overall success rate of approximately 60% may be expected (French 1993). There are some ethical concerns about the use of deception, although definitive diagnosis can allow patients to obtain proper therapy and avoid unnecessary antiepileptic drugs, with their side effects.

Video-EEG findings

During NESs, the EEG will show:

• no epileptiform activity
• no initial change such as premonitory spikes
• no postseizure slowing

Although the EEG tracing is frequently obscured by movement artifact, small interpretable segments containing alpha activity may be apparent, indicating that consciousness is preserved.

A normal or nonepileptiform EEG during a seizure may suggest a NES, but it can also occur during a simple partial seizure or frontal lobe complex partial seizure undetected by surface leads. A normal EEG during a seizure in which the patient is displaying generalized motor movements would not be expected in a true epileptic seizure, however.

The most important task is to ensure that the recorded event(s) are typical of the patient's spontaneous attacks. This task can be accomplished only by reviewing the recorded attack with a person who has witnessed such events. If it is determined that the recorded and spontaneous attacks are similar, a presumptive diagnosis of NESs can be made.

Some clinicians require that more than one attack be recorded, but this is not always possible. Nonetheless, it appears that a single recorded event similar to previous attacks is sufficient to consider NESs the most likely diagnosis.

This diagnosis, of course, does not exclude the possibility of coexisting epilepsy, especially if the patient has attacks with different clinical features. Some epilepsy patients experience psychogenic nonepileptic seizures at some point, and patients with psychogenic nonepileptic seizures can have neurologic illness.

The interictal EEG is not useful in making the distinction because it may be normal or abnormal in either case. The interictal EEG of patients with NESs may contain epileptiform discharges, even though the ictal record does not reveal electrographic seizure activity.

Alternatives to intensive video-EEG monitoring

If intensive video-EEG monitoring is not available, a diagnosis of NESs can be made with reasonable assurance using commonly available tools. Probably the best method is to obtain an EEG after the patient is sleep-deprived. A video camera can be set up in the EEG room. During the recording, and after explaining the procedure, apply techniques of suggestion, emphasizing the importance to the patient of recording an event.

The use of 24-hour ambulatory cassette EEG recording to diagnose NESs is not recommended unless a home video unit is available. Unless the behavioral aspects of the attack are recorded, there is too little diagnostic information. Moreover, excessive EEG artifacts during an attack often makes it very difficult to interpret the cassette EEG. If the patient has attacks characterized by staring with little motor activity, a cassette EEG can be useful. Certainly, differentiation of absence seizures from NESs characterized by loss of awareness is relatively easy. Again, simultaneous video recording greatly enhances diagnostic power.

Telling the patient about nonepileptic seizures

When the diagnosis of nonepileptic seizures (NESs) is confirmed, it must be presented in a positive, supportive, and nonthreatening manner. The physician must have an empathetic, compassionate attitude to ensure the patient's allegiance and offer hope for the future.

NESs should be described as a disorder that cannot be treated with anticonvulsant medication. Emphasize the disability associated with NESs and discuss the effect of the attacks on the patient's life. Also stress that NESs have psychosocial consequences as profound as those of epilepsy but point out that they do not require chronic treatment with drugs that produce side effects.

Outline the availability and indeed the success of treatment for NESs. Explain that many others suffer from NESs but that each person is unique, requiring tailor-made treatment. Reassure the patient that the outlook for improvement or complete recovery is excellent-even better than in the case of epilepsy.

Such a conference is time-consuming but rewarding. Most patients are willing to accept a diagnosis of NESs if it is presented in this manner, and they are eager to pursue an appropriate course of therapy.

What professionals should be involved in treatment?

Neurologists usually have little interest in continuing to care for patients in whom NESs are diagnosed, considering most of these patients not to have neurologic disease. Further, many psychiatrists are reluctant to care for patients with somatoform disorders. This reluctance is a problem for these patients, who suffer from significant disability. The psychiatrist should cooperate closely with the referring neurologist. The patient thus gains by receiving continuing support on the medical side while exploring the root causes of the psychiatric disability.

Probably the best approach to treatment for most patients is offered by a multidisciplinary team composed of specialists in the fields of

• neurology
• psychiatry
• psychology
• social work

Such a group addresses the multidimensional nature of NESs. Besides diagnosing and managing the underlying psychiatric condition, the treatment can also address problems of psychosocial functioning, occupational issues, and family interactions. Coordination of efforts in these several spheres is key to the team approach. The goal is to restore the patient to normal functioning in the shortest possible time.

Types of programs

The patient can be treated as either an inpatient or an outpatient. The goals for both approaches are the same and in the long run they probably produce similar outcomes.

Inpatient programs are usually associated with epilepsy centers. Two excellent ones are located in Minneapolis and St. Paul, Minnesota. On discharge, the patient is referred to a local psychiatrist, who carries out longer-term intervention. The major advantage to the inpatient approach is the intensity and focus that is possible at the onset of treatment. A major disadvantage is the cost associated with hospitalization.

The outpatient approach is less intense but cost-effective. The outpatient approach combines regular team meetings with ongoing therapy. The principal therapist may be the psychiatrist, psychologist, or social worker, depending on the nature and extent of the patient's particular problem.

Each patient with NESs requires an individualized treatment program. Those with significant depression may be treated with antidepressants along with ongoing psychotherapy. Patients with panic disorder may receive appropriate pharmacotherapy with an agent such as desipramine. Some will require only short-term psychotherapy.

For patients with somatoform disorder, supportive treatment and encouragement can aid overall functioning and stop "doctor shopping" for multiple somatic complaints. NESs are likely to continue in this group, however.

Prognosis

The prognosis for patients with NESs is variable. Even without specific treatment, NESs tend to decline over time. A 1-year follow-up study of 80 patients in whom NESs were diagnosed during EEG-video monitoring revealed that 59% had become seizure-free in the interval, and an additional 20% reported a decline in event frequency. Other studies have found that even with specific intervention, fewer than 50% become seizure-free, and some who become seizure-free develop other (substitution) conversion symptoms.

The literature is sparse with respect to the effectiveness of psychotherapy, but some improvement occurs in up to 60% of patients. Success rates appear to depend on the severity and duration of the psychiatric illness. Those with NESs of relatively short duration, perhaps related to an intercurrent situation such as an acute loss, have a favorable prognosis. Short-term psychotherapy may be highly effective in these cases.

On the other hand, prolonged duration of NESs associated with severe psychopathology may be refractory to treatment. In these cases, supportive treatment may be all that can be offered.

An optimistic attitude on the part of the caregivers is important and indeed justified. This attitude should be conveyed to the patient at the outset and continually reinforced.

Dodrill CB, Wilkus RJ, Batzel LW. The MMPI as a diagnostic tool in non-epileptic seizures. In: Rowan AJ, Gates JR, eds. Non-epileptic seizures. Boston: Butterworth-Heinemann; 1993, pgs 211-219.

Ettinger AB, Dhoon A, Weisbrot DM, Devinsky O. Outcome in psychogenic nonepileptic seizures: 1 to 10-year follow-up in 164 patients. Ann Neurol. 2003;53:305-11.

Ettinger AB, Dhoon A, Weisbrot DM, Devinsky O. Predictive factors for outcome of nonepileptic seizures after diagnosis. J Neuropsychiatry Clin Neurosci. 1999;11:458-63.

French JA. The use of suggestion as a provocative test in the diagnosis of psychogenic non-epileptic seizures. In: Rowan AJ, Gates JR, eds. Non-epileptic seizures. Boston: Butterworth-Heinemann; 1993, pgs 101-109.

Pritchard PB. The role of prolactin in the diagnosis of non-epileptic seizures. In: Rowan AJ, Gates JR, eds. Non-epileptic seizures. Boston: Butterworth-Heinemann; 1993, pgs 93-100.

Ramsay RD, Cohen A, Brown MC. Coexisting epilepsy and non-epileptic seizures. In: Rowan AJ, Gates JR, eds. Non-epileptic seizures. Boston: Butterworth-Heinemann; 1993, pgs 47-54.

Reuber M, Pukrop R, Bauer J, et al. Psychogenic nonepileptic seizures: review and update. Epilepsy Behav. 2003;4(3):205-16.

What is status epilepticus?

A patient who has continuous seizures or does not recover between recurrent seizures that are "so frequently repeated or so prolonged as to create a fixed and lasting condition" is considered to have status epilepticus (SE). Clinical and experimental data indicate that seizure activity for 30 minutes is a reasonable criterion for use of the term, at least for recurrent seizures, because brain damage is likely to occur at this point. Aggressive treatment should start much sooner, however.

Are there different types?

Convulsive status epilepticus (SE) is often associated with poor outcome. This section presents the etiology, related electrophysiology, course and consequences of convulsive SE.

Important clinical lessons about the effects and prognosis of convulsive SE have emerged from studies:

• Some damage appears to accrue from the epileptic neural activity itself, although systemic factors such as hypotension, hypoxia, and acidosis may add to the neurologic complications.
• The longer the duration of SE, the more refractory it becomes and the more neuronal damage occurs.
• Clinical and experimental data suggest that 30 minutes of convulsive SE is a critical duration in neurologic function and probably in prognosis. This figure is far less certain for other forms of SE, particularly nonconvulsive forms.

Complications

Generalized convulsive status epilepticus (GCSE) is the most dramatic, most dangerous, and best-studied type of SE. It is potentially life-threatening but also treatable. A plan for medical management and pharmacotherapy is crucial. The clinician needs to understand its etiology, electrophysiology, pathophysiology, course, and consequences.

Symptoms

Convulsive SE is readily recognizable. It may start with simple or complex partial seizures but often begins with a generalized convulsion. Convulsions recur, most lasting only a minute or so, along with intervals of persistent unresponsiveness. Each convulsion may begin with several seconds of a tonic phase with tensing of extensor muscles and forced expiration, followed by a clonic phase with gradually slowing clonic movements. Both phases usually involve bilateral and symmetric movements, although there may be a focal onset with head or eye deviation, even without unilateral limb movement. Consciousness is impaired, at least from the time of tonic seizures.

Less often, convulsions are continuous. In this case clonic movements eventually diminish, often being replaced by repetitive jerking movements of the eyes, eyelids, or facial muscles alone or sometimes with intermittent limb jerking. These signs constitute "subtle" SE and imply continuing epileptic brain activity with a "decoupling" of electrical and motor systems.

Incidence

The incidence of convulsive SE is usually estimated to be about 60,000 cases each year in the United States (probably over half of them in children), but population-based surveys suggest that it may occur several times as often. The incidence of other forms of SE is less well documented.

Etiology and epidemiology

Convulsive SE is not a disease itself but, rather, a serious manifestation of some underlying disorder. The following list shows etiologies and percentages of patients affected; these figures are obtained from a summary of several studies of adult patients.

These percentages total more than 100% because of multiple causes. For instance, a patient with a congenital lesion and chronic epilepsy may experience anticonvulsant withdrawal or infection.

The causes of convulsive SE may vary tremendously in different populations. In urban hospitals, for example, SE is more often related to alcohol and drugs. The causes or precipitants of convulsive SE are also different in patients with known epilepsy than in those presenting with acute, new illness. Congenital abnormalities and infection increase in importance in children.

Often, there is an interaction between acute systemic illness and earlier neurologic disease, including epilepsy and other earlier neurologic insults. A history of epilepsy is often assumed, but in actuality about two-thirds of SE cases occur in patients who have not had prior seizures.

About 1% of patients with epilepsy will have an episode of SE in a given year. Anticonvulsant withdrawal is often assumed in patients with epilepsy, although this may be less frequent than presumed. Anticonvulsant changes initiated by physicians may cause withdrawal seizures as often as patient noncompliance. Adding new anticonvulsants may alter metabolism and lead to subtherapeutic or toxic levels of prior medications.

Infections may have a role in epileptogenicity, but several antibiotics also can precipitate seizures and alter anticonvulsant metabolism.

The epidemiology of convulsive SE has several clinical implications:

• Convulsive SE usually has an identifiable cause. Look for it. Trauma, new or prior vascular disease, metabolic derangements, drug toxicity (due to prescribed or "recreational" drugs), and infection not only help to explain the SE but often determine the subsequent course; they must be found to be treated appropriately. Alcohol abuse and benzodiazepine withdrawal are common contributors.
• There is often more than one cause or precipitant: medication withdrawal, infection, or sleep deprivation may add to an earlier illness and precipitate convulsive SE. In some series, up to 50% of patients have either an infection or a recent medication change. Conversely, even in acute illness, convulsive SE occurs more often in people with prior neurologic deficits.
• Convulsive SE can be the first sign of neurologic disease, especially in children, in whom up to 10% of initial seizures (particularly febrile seizures) may be SE.

Convulsive status epilepticus (SE) is rarely difficult to diagnose. Other forms of SE, however, may be mistaken for altered mental status from other causes, or for movement disorders. These other forms include both generalized seizures and partial or focal seizures.

The management of other forms of SE is often similar, though less urgent, than the management of convulsive SE.

Absence (petit mal) SE

Absence SE implies generalized epileptiform EEG discharges. Synonymous (and somewhat confusing) terms include "spike wave stupor" and "epileptic twilight states."

Because of the difficulty of detection, it is hard to know how often absence SE occurs. It may occur one-fourth as often as convulsive SE. Three different syndromes have been described:

• Classic absence SE typically occurs in younger patients with absence epilepsy. Very frequent classic absences or (less frequently) continuous absence is noted.



• Atypical absence SE occurs in children with symptomatic generalized epilepsy (e.g., the Lennox-Gastaut Syndrome). There is decline of mental status associated with significantly increased persistence of generalized slow spike wave. Diagnosis may be difficult because cognitive status is often impaired at baseline in this population and generalized slow spike wave is quite common in the baseline EEG as well.



• Absence SE of late onset occurs in a variety of circumstances. Some patients may have had previous symptomatic epilepsy but have been seizure-free and off medications for years. In other older patients, absence SE is the first presentation of epilepsy. It may be precipitated by medications such as Cipro or baclofen, or may occur following generalized convulsions or electroconvulsive therapy. Misdiagnosis is common unless there is a high level of suspicion; decreased mental status is often attributed to psychiatric or other systemic illnesses.

Onset may be sudden or gradual. Patients may be awake, walking and talking, although they are often confused. Motor activity may be preserved but clumsy. There is occasionally some blinking or myoclonus.

Absence SE can persist for days or weeks without being recognized. A history of epilepsy is suggestive. Most cases may be missed, but EEG can readily confirm the diagnosis. Some EEGs show generalized 3-Hz epileptiform discharges. Some may be secondarily generalized from a focus.

EEG of a 33-year-old woman with a history of epilepsy, now ambulatory and speaking, but confused at the time of office visit; approximately 3-HZ generalized spike-and-slow wave discharges with a frontal and central emphasis.

Especially in younger patients with primary generalized epilepsies, benzodiazepine therapy is often rapidly successful. Valproic acid may be better at preventing recurrences. Most patients return to normal, although recurrence of absence SE is relatively common. In older patients and those whose SE has a less certain cause, the response to anticonvulsant therapy can be delayed.

The typical absence status of patients with prior epilepsy is not thought to be life-threatening, but occasionally episodes end with a generalized convulsion. Absence SE is probably the most benign type of SE in terms of potential neuropathophysiologic consequences and may warrant less aggressive treatment than other forms of SE. For example, treatment with anesthetic drips is usually not pursued in this situation.

Myoclonic status epilepticus

Myoclonic status epilepticus (MSE) also occurs in several forms. The most severe form, which is highly associated with the eventual death of the patient, occurs after anoxia. Persistent myoclonus due to a severe encephalopathy without MSE is potentially reversible. The EEG helps determine whether myoclonus is the sporadic sign of an encephalopathy (with a better prognosis) or part of MSE. After anoxia, MSE is better considered a sign of a severely damaged brain than a treatable epileptic condition. MSE may have motor manifestations limited to "subtle" status, also an ominous sign after anoxia.

Less severe forms of MSE occur as a manifestation of generalized epilepsies such as juvenile myoclonic epilepsy. This form of MSE may include prolonged epileptic myoclonus without loss of consciousness. The EEG shows very characteristic generalized polyspike and slow wave discharges with a normal background between episodes. Episodes may go on for hours (with preserved consciousness), but the prognosis is excellent given the prior normal neurologic function.

Patients can also return to baseline in the MSE of progressive myoclonus epilepsies, although the epilepsy may be part of a progressive debilitating disease.

In all these conditions, the EEG can help distinguish MSE from encephalopathies with less rhythmic abnormalities.

Tonic or clonic SE

Tonic SE is rare. It is seen primarily in children, particularly those with Lennox-Gastaut syndrome. This form of SE does not respond well to medications, as is true of most seizure types in Lennox-Gastaut syndrome. Rarely, benzodiazepines have been cited as triggering tonic status. Tonic (and atonic) SE is distinctly uncommon in adults, especially those with normal interictal neurologic function.

Generalized clonic SE is also a pediatric condition. Clonus is often of low amplitude. Both sides of the body are usually involved but may move asynchronously.

Complex partial SE

Complex partial status epilepticus (CPSE) implies impairment of consciousness due to seizures with a focal cortical origin. The disorder is sometimes called fugue state or psychomotor status.

Confusion is the most common symptom. A sudden alteration in behavior, particularly in a patient with prior epilepsy, should suggest the possible diagnosis of CPSE. A diagnosis of CPSE is frequently invoked to explain bizarre behavior, but this type of SE is actually uncommon. Some patients may have complex partial seizures with a prolonged postictal phase. CPSE may be continuous or include frequent discrete seizures without recovery between them; the latter series of spells account for some of the prolonged episodes.

CPSE can go on for weeks or even months, and patients may have prolonged cognitive deficits after CPSE. It can be very difficult to distinguish from absence SE. Patients with CPSE may exhibit more bizarre behavior during seizures, so the disorder also is confused with psychiatric disease or metabolic encephalopathies with delirium.

The usual site of origin is assumed to be mesial temporal structures with limbic connections, but recordings obtained from implanted electrodes have shown that frontal lobe onset is common. The EEG may show the seizure clearly (but this is less likely in frontal areas without implanted electrodes) or may show just persistent focal slowing. Seizures may spread rapidly, and nonconvulsive SE with generalized discharges ("absence SE") may actually arise from a focus.

Increasing reports of cognitive and other sequelae following prolonged CPSE have lent a greater urgency to its interruption, but treatment is usually not as aggressive as for generalized convulsive SE. Most CPSE responds relatively rapidly to anticonvulsants but may recur frequently and even regularly in the same patient, even with adequate anticonvulsant therapy.

Simple partial SE (epilepsia partialis continua)

Epilepsia partialis continua (EPC) usually refers to partial motor status. Continuous unilateral facial or hand clonic jerking are most commonly observed, but any group of muscles can be involved. An acute structural lesion must be considered and ruled out, but lesions are not always found even with contemporary imaging. Consciousness may be impaired because of secondary effects of the structural lesion. The EEG often demonstrates periodic lateralizing epileptiform discharges (PLEDs) in a corresponding distribution.

As in all other areas of medicine, effective treatment of status epilepticus (SE) is facilitated tremendously by the correct diagnosis! Convulsive SE is rarely a diagnostic difficulty, but nonconvulsive forms, including episodes after generalized seizures, may be difficult to recognize or missed altogether.

Differential diagnosis

Almost all published protocols and guidelines refer to generalized convulsive SE (GCSE). Generalized nonconvulsive SE (NCSE) after convulsions should probably be considered as much of an emergency as GCSE. Other forms of SE are of less certain morbidity and urgency. Medication use is generally similar, if less immediate. Nevertheless, the pathophysiologic underpinnings of many different types of SE (with the possible exception of absence SE, EPC, and myoclonic status after anoxia) argue for urgent treatment in almost all cases. Other forms of SE can lead to convulsions, and a casual approach is inappropriate.

Phenytoin and now fosphenytoin are probably the most frequently used anticonvulsants for SE. The older formulation (phenytoin in a basic solution including ethylene glycol) is still used because its cost is substantially lower. However, intravenous phenytoin has three major disadvantages:

• A longer time is required to achieve therapeutic levels.
• Intramuscular administration is not possible.
• Significant soft tissue damage may occur with extravasation.

The newer formulation (fosphenytoin, a phenytoin prodrug) can be administered more quickly, an important advantage when seizures must be controlled quickly. Because a basic solution is not required for dissolving fosphenytoin, intramuscular administration is safe and extravasation carries little risk.

Cardiac rhythm abnormalities and hypotension can occur with both formulations.

Both phenytoin and fosphenytoin cause minimal sedation. This advantage over other anticonvulsants is probably overstated but it may be pertinent for patients with head trauma, hemorrhage, or raised intracranial pressure, for whom it is important to monitor alertness. Patients with GCSE are unconscious, and the most immediate concern is stopping the seizures.

In the absence of acute structural lesions, both phenytoin and fosphenytoin may be successful alone in up to 80% of patients with GCSE. If either is used, oral phenytoin can then become a long-term maintenance medication. This obviates the necessity of changing or adding drugs.

Patients may need adjunctive benzodiazepines to interrupt convulsions if they occur during phenytoin infusion. The VA Cooperative Study, for instance, found that phenytoin with diazepam stopped GCSE more quickly than phenytoin alone. Many authorities recommend phenytoin as the primary treatment of GCSE, sometimes after initial interruption of convulsions with a benzodiazepine.

A usual loading dose for phenytoin is 15 mg/kg, but 20 mg/kg is reasonable before concluding that phenytoin is insufficient. It should be given by intravenous bolus or in saline solution at a maximum rate of 50 mg/min. (It may precipitate in glucose solutions.) Intramuscular phenytoin is poorly absorbed and should not be used.

The usual loading dose for fosphenytoin is 15-20 mg phenytoin equivalent (PE)/kg. It can be administered as quickly as 150 mg PE/min and there are theoretical advantages to doing so. Rapid rates are associated with greater risk of cardiac toxicity with both formulations, however. Intramuscular fosphenytoin is safe and readily absorbed; a "loading dose" typically leads to therapeutic levels within half an hour.

Conduction defects are the primary cardiac toxicity, but hypotension is not rare. Cardiac monitoring is appropriate during phenytoin infusion. Elderly patients or those with cardiac disease may not tolerate phenytoin as well as phenobarbital. Acute toxicity is more closely related to the infusion rate than to total dose. Patients with possible complications may tolerate greater doses in slower infusions.

The recently available intravenous formulation of valproic acid has several advantages:

• It is not sedating
• Hypotension and respiratory comprise are not issues
• It can be administered relatively quickly

It has not been compared directly to phenytoin, fosphenytoin, or phenobarbital, however, and may not be as effective in GCSE.

Valproic acid is administered in a loading dose of 25 mg/kg. Higher doses are occasionally used. The package insert recommends an infusion rate of no faster than 20 mg/min, but several reports have documented that infusion rates between 100 and 150 mg/min are safe and well tolerated.

Valproic acid may significantly increase phenobarbital levels or free phenytoin levels if these medications have already been administered.

Caution should be exercised if the patient is being treated with lamotrigine (Lamictal, a widely used maintenance anticonvulsant) because administration of valproate may quickly double lamotrigine levels. It is prudent to hold lamotrigine doses in this situation and check serum levels within 24 hours.

Rectally administered valproic acid syrup is absorbed irregularly and is rarely used now that an intravenous formulation is widely available. It can be useful if intravenous access can not be obtained, however, because it is absorbed more rapidly than valproic acid given by mouth and bypasses the hepatic first-pass effect. The usual dose is 25 mg/kg or higher. Valproic acid syrup is cathartic. It should be mixed in equal volumes of water prior to administration.

Phenobarbital is often used if phenytoin is insufficient, but it is frequently avoided out of fear it will cause sedation or respiratory compromise (especially after several doses of benzodiazepines). One of its advantages is a relative lack of cardiac toxicity until very high doses are reached.

Phenobarbital has been compared favorably with a combination of diazepam and phenytoin; in the VA Cooperative Trial, clinical response was actually faster with phenobarbital. Respiratory depression and hypotension are likely if phenobarbital, a barbiturate, is used together with benzodiazepines.

Loading with up to 20 mg/kg is reasonable. It may be administered as quickly as 100 mg per minute in normal-size adults if respirations are carefully monitored. Loading is faster than with phenytoin, but its lipid solubility is lower and brain penetration is slower. Nevertheless, phenobarbital may act quickly, even before therapeutic levels are established.

Some SE may be refractory to phenytoin, but high enough doses of phenobarbital will control almost all seizures. Very high doses require artificial ventilation and may cause hypotension, but they may be tolerated better than expected. Sedation must be expected with high doses, but levels below 40 mcg/mL should not produce prolonged coma.

Diazepam (Valium, Diastat) is a very common first treatment for SE, but conservative management would restrict it to patients with continuing convulsions or those having another convulsion during infusion of a maintenance medication. Reported efficacy rates in SE have varied widely, from 38% to 83%. Diazepam can interrupt convulsive SE rapidly but should not be used alone.

The usual practice is to administer 10 mg (0.15 mg/kg) intravenously over a few minutes, repeating if necessary. Rectal administration has been effective, particularly in children, and a gel suitable for rectal administration (Diastat) is available. Doses of the rectal gel necessary to terminate seizures are somewhat higher and vary by weight and age. A buccal formulation (Diazapam Intensol) is less well studied but widely considered effective as well. Both rectal and buccal formulations bypass first-pass hepatic metabolism and thus result in reasonable levels within about 30 minutes. In contrast, both intramuscular and oral diazepam are absorbed slowly and should not be used when acute seizure control is necessary.

Diazepam is very lipid-soluble, enters the brain rapidly, and may have an anticonvulsant effect within a minute after intravenous administration. Nevertheless, it redistributes to many tissues and its CNS effect declines in 20 to 30 minutes, so longer-acting anticonvulsants should be used concomitantly to prevent recurrent seizures or SE. Repeated doses may lose effectiveness but produce metabolites with prolonged elimination half-lives and potential toxicity, including prolonged coma.

Continuous infusion of diazepam (generally 4 to 8 mg/h) is often discussed for the management of SE but is rarely practiced, probably because the optimal doses have not been clearly established and rapid acute tolerance may develop. Continuous infusion should be used in intensive care units only. Iatrogenic apnea (often ascribed to seizures or to "tongue swallowing") can occur suddenly.

Lorazepam (Ativan) has several advantages:

• It acts rapidly enough to interrupt seizures quickly.
• It also has a more prolonged anticonvulsant effect. (Some physicians consider it to satisfy the requirements for both the acute interruption of SE and prolonged protection against recurrence.)
• Compared to diazepam, doses are approximately half as large and it is half as lipid-soluble. Its brain penetration is slower but still rapid.
• Its effect declines less rapidly than diazepam.
• It has no active, troublesome metabolites.

Many epileptologists prefer lorazepam because of its favorable pharmacokinetics, but direct comparative studies are few. A double-blind, randomized trial found lorazepam marginally more effective than diazepam in controlling SE, with an onset of action not significantly different. Adverse effects are similar to those of diazepam, although perhaps less sudden. Lorazepam may provide 12 hours of anticonvulsant effect, but acute tolerance may reduce its maintenance value.

Midazolam (Versed) is used both acutely to treat SE at presentation and as a continuous anesthetic drip after SE has proven refractory to standard treatment (see below). Acutely, it may be given in intravenous boluses of 0.2 mg/kg (5 to 20 mg). Onset of therapeutic effect is extremely rapid because of high lipid solubility, but its effect is short-lived, and relapses of seizures may be expected. The very short duration of action allows clinical assessment soon after its discontinuation.

Midazolam may be the best intramuscular treatment of SE when this is the only available route.

Clonazepam (Klonopin) appears similar to other benzodiazepines and is popular in Europe, but it is not available in intravenous form in the United States for the treatment of SE.

When a benzodiazepine and one or more longer-acting anticonvulsants have not terminated SE, strong consideration should be given to treatment with continuous intravenous or inhalational anesthesia. This always requires intubation, respiratory support, and treatments to maintain normal blood pressure. Continuous EEG monitoring is mandatory in this situation.

The goal is escalation of anesthetic medications until seizures are controlled or side effects (usually hypotension) preclude further dose increases. Many have suggested that if SE has been difficult to control, anesthetic medications should be increased until a "burst suppression" or flat record is achieved. This state is then maintained for at least 24 hours and the anesthetic medications are gradually tapered. Maintenance anticonvulsant medications are administered in the interim (usually by nasogastric tube) in the hope that these will prevent recurrence of seizures as the anesthetic medications are tapered.

Because three longer-acting anticonvulsants are now available in intravenous form, it is tempting to try all three before moving on to intubation and anesthetic treatment. In GCSE, this may be the wrong thing to do. In the VA Cooperative Study, only 7% of patients failing the first treatment responded to the second and only 2% of those failing two treatments responded to a third. A strong argument can therefore be made for proceeding to continuous anesthetic treatment within a half-hour of onset of GCSE.

Midazolam

Midazolam, which can be used for acute treatment at presentation (above), can also be used as a continuous anesthetic treatment for refractory SE. A loading dose of 0.2 mg/kg is typically administered, followed by a drip of 0.2 to 2 mg/kg per hour. Achieving burst suppression or isoelectric EEG typically requires escalation towards the higher doses. Hypotension appears less problematic than with the other anesthetics discussed. Tachyphylaxis may occur after 48 hours and require further escalation of doses.

The very short half life allows rapid assessment of mental status after the drip is tapered and terminated, a decided advantage over pentobarbital. The more rapid recovery may also shorten time intubated or in the ICU. Cost was a concern until the generic intravenous formulation became available.

Propofol

A 2mg/kg loading dose is typically administered followed by a 5 to 10 mg/kg per hour drip. Hypotension can occur. The high lipid content of the formulation is a concern, especially in patients with significant hyperlipidemia or atherosclerosis. Involuntary myoclonic movements have been reported at induction during routine surgeries but their meaning is uncertain and we have not encountered them in the treatment of SE. The medication is very lipid-soluble and has a short half life, allowing a more rapid recovery after tapering than with pentobarbital.

Pentobarbital

Loading doses of 3 to 5 mg/kg followed by infusion of 1 to 4 mg/kg per hour are typical.
Effectiveness is assessed by effect on the EEG, with an attempt to eliminate seizures or aim for burst suppression. (Most authors seek a burst suppression pattern.) Blood levels are more useful for indicating residual toxicity than for assessing therapeutic effect.

The half-life of pentobarbital is approximately 20 hours (shorter than for phenobarbital, so it dissipates sooner), but it may be extended at higher levels. Pentobarbital accumulates in fatty tissues other than brain with prolonged treatment and these stores must be mobilized and secreted after administration ceases. One should therefore not attribute prolonged coma after pentobarbital treatment to a "burnt-out" brain before the medication has had time to dissipate.

All SE should be suppressible with adequate pentobarbital doses, but hypotension is common. Usually, volume replacement and low doses of vasopressors are sufficient. Myocardial function and temperature regulation can be impaired. Most reports of pentobarbital use show a very high mortality, usually attributed to severe underlying diseases causing SE refractory enough to require pentobarbital.

An advantage of pentobarbital, besides its invariable effectiveness when used in large enough doses, is that it reduces cerebral metabolism and blood flow. The infusion is also easy to adjust. The optimal duration of barbiturate-induced coma has not been established; recommendations range from 4 to 72 hours. Probably pentobarbital should not be withdrawn until the patient has a therapeutic blood level of two other anticonvulsants.

Inhaled anesthetics

Inhaled anesthetics are less well studied and far less convenient than the drugs already described but may be useful for patients who are allergic to pentobarbital. Most inhaled anesthetics increase cerebral blood flow, a theoretical disadvantage. This problem probably applies least to isoflurane, possibly the most effective anesthetic with the least cardiovascular effect in the setting of SE. Halothane is used relatively frequently, but isoflurane may produce a burst suppression EEG tracing with less severe cardiovascular morbidity. Vasopressors may be needed with both agents. Nitrous oxide does not appear effective. Enflurane can precipitate convulsions.

Other agents

Neuromuscular blocking agents eliminate motor activity, but they are not anticonvulsants. They may provide false reassurance when SE continues on an electrical and metabolic basis. They can help when excessive movement impairs oxygenation, acid-base balance, or temperature regulation, but adequate doses of anticonvulsants, particularly pentobarbital, will obviate these problems.

Steroids and osmotic agents may be used to treat cerebral edema that results from prolonged SE, particularly in children, but their efficacy has not been established.

Rarely, a persistent seizure focus causing refractory partial SE may be resected surgically.

Most medication trials have studied patients with GCSE. The same medications may treat other forms of SE. Intravenous benzodiazepines usually interrupt absence SE, and subsequent treatment may be unnecessary. For patients with continuous generalized discharges and coma, however, intravenous benzodiazepines are often insufficient.

For partial or nonconvulsive SE, enteral valproate and carbamazepine are more valuable, although the response may take days; rarely will pentobarbital or anesthetics be necessary. In children, phenobarbital is often preferred to phenytoin because of better absorption, greater efficacy, and possibly fewer long-term side effects.

When SE does not respond to treatment as expected, the clinician's attention should refocus along several lines:

• Is the diagnosis of SE correct?
• Has the underlying cause been correctly assessed? This is crucial, because SE is most likely to continue when trauma, hemorrhage, or infections such as encephalitis remain untreated.
• Have medications been given in adequate doses? (The 1000-mg standard phenytoin infusion may be insufficient, for example.)
• Have medications been adequately absorbed? Absorption can be a problem if there are difficulties with intravenous access or if the drug is given by another route.
• Has the SE recurred after successful treatment? This situation most often results from inadequate attention to maintenance levels of longer-acting anticonvulsants or lack of treatment of the underlying disease.

Alldredge BK, Gelb AM, Isaacs SM, et al. A comparison of lorazepam, diazepam, and placebo for the treatment of out-of-hospital status epilpeticus. N Engl J Med 2001;345:631-637.

Barry E, Hauser WA. Status epilepticus: the interaction of epilepsy and acute brain disease. Neurology. 1993;43:1473-1478.

Classen J, Hirsh LJ, Emerson RG, Mayer SA. Treatment of refractory status epilepticus with pentobarbital, propofol, or midazolam: a systematic review. Epilepsia 2002;43:146-153.

Engel J Jr. Status epilepticus. In: Seizures and Epilepsy. Philadelphia: FA Davis; 1989:256-280.

Leppik IE, Derivan AT, Homan RW, et al. Double-blind study of lorazepam and diazepam in status epilepticus. JAMA 1983;249:1452-1454.

Lothman EW. The biochemical basis and pathophysiology of status epilepticus. Neurology. 1990;40(suppl 2):13-23.

Lowenstein DH, Alldredge BK. Status epilepticus at an urban public hospital in the 1980s. Neurology. 1993;43:483-488.

Lowenstein DH, Alldredge BK. Current concepts: Status epilepticus. N Engl J Med 1998;334:970-976.

Meldrum BS, Horton RW. Physiology of status epilepticus in primates. Arch Neurol. 1973;28:1-9.

Tomson T, Lindbom U, Nilsson BY. Nonconvulsive status epilepticus in adults: thirty-two consecutive patients from a general hospital population. Epilepsia. 1992;33:829-835.

Treiman DM, Meyers PD, Walton NY, et al. A comparison of four treatments for generalized convulsive status epilepticus. N Engl J Med 1998;339:792-798.