Refractory status epilepticus in adults

INTRODUCTION — Status epilepticus is a medical and neurologic emergency that requires prompt evaluation and treatment. This topic will review refractory status epilepticus (RSE).

The definition, classification, clinical features, diagnosis, and initial treatment of convulsive status epilepticus in adults are reviewed separately. (See "Convulsive status epilepticus in adults: Classification, clinical features, and diagnosis" and "Convulsive status epilepticus in adults: Management".)

Nonconvulsive status epilepticus and the diagnosis and management of status epilepticus in children are discussed elsewhere. (See "Nonconvulsive status epilepticus: Classification, clinical features, and diagnosis" and "Clinical features and complications of status epilepticus in children" and "Management of convulsive status epilepticus in children".)

DEFINITION AND ETIOLOGY

Refractory status epilepticus — Refractory status occurs in approximately 20 percent of patients with status epilepticus [1-5]. Longer seizures are less likely to stop spontaneously and are also less responsive to antiseizure medications [6,7].

●Definition – The International League Against Epilepsy (ILAE) follows the operational definition of generalized convulsive status epilepticus as having a duration of five minutes [8].

RSE, however, is not defined by duration, but rather as status epilepticus that does not cease with administration of two antiseizure medications (administered in appropriate and adequate doses; usually an intravenous benzodiazepine, followed, if necessary, by a longer-acting antiseizure medication). RSE occurs in approximately one-quarter of status epilepticus cases and has higher morbidity and mortality [5].

●Etiologies – In a global registry that collected information from 2013 to 2017 for 776 patients with RSE, the most common underlying etiologies were classified as cryptogenic (26 percent), infections including acute encephalitis and meningitis (19.6 percent), vascular including stroke (14.5 percent), anoxic (11.1 percent), antiseizure medication reduction or withdrawal (7.3 percent), cerebral tumor (6.1 percent), and trauma (5.5 percent) [9].

Super-refractory status epilepticus — Super-refractory status epilepticus (SRSE) is defined as status epilepticus persisting or recurring after 24 hours or more of treatment with highly sedating antiseizure medications (eg, midazolam, pentobarbital, or propofol; sometimes referred to as anesthetics) or when therapy is tapered after 24 hours of use [10]. A 2023 systematic review and meta-analysis identified studies reporting 1200 patients (mean age 41 years) with nonanoxic SRSE [11]. The mean duration of SRSE was 36 days, and the most common etiologic categories were acute cerebral events and unknown etiologies (42 and 22 percent, respectively). Successful treatment (cessation of SRSE) was achieved for 81 percent of patients, but the in-hospital mortality rate was 24 percent and the rate of no to moderate disability at discharge was only 27 percent.

New-onset refractory status epilepticus — New-onset refractory status epilepticus (NORSE) is a syndrome described in several reports of adults and children who present with severe, treatment-refractory generalized seizures and status epilepticus of unclear etiology, often in the setting of a prodromal febrile illness suggesting a viral encephalitis [12-17]. The term NORSE applies to less than 10 percent of patients presenting with status epilepticus but a higher proportion of patients with refractory nonconvulsive status epilepticus [18,19]. (See "Nonconvulsive status epilepticus: Classification, clinical features, and diagnosis".)

An expert panel has defined NORSE as a specific clinical presentation (but not a specific diagnosis) of new-onset RSE without a clear acute or active structural, toxic, or metabolic cause, occurring in a patient without earlier epilepsy or another neurologic disorder known to precipitate seizures [20]. NORSE refers to RSE for which no such etiology is identified within the first 72 hours after presentation, although experts in the field include cases of viral encephalitis or autoimmune encephalopathy even if diagnosed within 72 hours [20,21].  

A cause for NORSE (table 1) is ultimately found in approximately one-half of patients, most frequently an autoimmune or paraneoplastic encephalitis (eg, NORSE due to anti-N-methyl-D-aspartate receptor encephalitis or LGI-1 autoimmune encephalopathy). When no underlying cause is found, the situation is referred to as cryptogenic NORSE or NORSE of unknown etiology. In a 2023 systematic review and meta-analysis that included 109 studies with 395 adult cases of NORSE, the etiology (identified after a negative initial evaluation) was cryptogenic in approximately 50 percent of cases, autoimmune in 36 percent, and infectious in 8 percent [22].

A similar syndrome in children termed febrile infection-related epilepsy syndrome (FIRES) is characterized by a prodromal febrile infection starting between two weeks and 24 hours before the onset of RSE, with or without fever at onset of status epilepticus. FIRES is considered a subcategory of NORSE. (See "Clinical features and complications of status epilepticus in children", section on 'FIRES and NORSE'.)

APPROACH TO TREATMENT

Neurology consultation and intensive care

●Generalized RSE – For patients with generalized convulsive RSE, essential management steps (if not already done) include emergency consultation with neurology, transfer to an intensive care unit with continuous electroencephalography (EEG) capability, and rapid sequence intubation and mechanical ventilation [1].

Continuous EEG monitoring should be instituted as soon as possible, along with continuous pulse oximetry and blood pressure monitoring, often with an arterial catheter. Vasopressors should be available at the bedside. (See "Use of vasopressors and inotropes".)

These measures are necessary because pharmacologic therapy for RSE employs highly sedating medications (eg, midazolam, propofol, pentobarbital) at doses that can induce coma. Regardless of the specifics of pharmacologic therapy, it is critical to provide adequate ventilatory and hemodynamic support.

●Focal RSE – For patients with focal RSE, a less aggressive approach is reasonable; treatment with conventional antiseizure medications should be exhausted before using highly sedating medications, particularly in nonintubated patients. (See "Convulsive status epilepticus in adults: Management", section on 'Focal motor status epilepticus'.)

Continuous EEG monitoring — Continuous EEG monitoring is critical during the treatment of RSE. Portable EEG monitoring should be started in the emergency department (when available) if rapid transfer to an intensive care unit is delayed. Once continuous infusion of midazolam, pentobarbital, or propofol has begun, continuous EEG monitoring is necessary to confirm that seizures have been treated adequately; to guide use of antiseizure medications and assess the level of suppression achieved; and to monitor for relapse of seizures and status epilepticus, especially when infusions are tapered.

Infusion therapies and other treatments

Choice of infusion agent — The primary drugs used for generalized RSE are midazolam, propofol, pentobarbital (or thiopental in some countries), and sometimes, ketamine [23]. We suggest initial therapy with midazolam. Many experts prefer to start with midazolam because of a better safety profile at high doses. Although comparative data are limited, observational evidence suggests that midazolam and propofol are equally effective in controlling RSE [24]. Midazolam or propofol are preferable to barbiturates because these drugs offer the possibility of a quick resolution of the status epilepticus with a shorter duration of sedation. This can be particularly helpful for patients who are at risk for ventilator dependence with prolonged therapy (eg, those with severe pulmonary disease, severe debilitation, or malignancy). Careful hemodynamic monitoring is important in hypovolemic patients given propofol, which can cause hypotension.

Treatment with high-dose pentobarbital (or thiopental, used more in Europe) is still in use because of the greater experience with its use, and because it appears to offer longer-lasting seizure control than alternative drugs [23].

Longer infusions and higher doses of propofol may precipitate the propofol infusion syndrome (see 'Propofol' below). Both barbiturates and propofol may worsen hemodynamic problems in unstable patients; the primary alternative, midazolam infusion, is usually well tolerated in this setting, although it can also cause hypotension at higher doses.

Factors to consider when selecting pharmacologic therapy for RSE include the urgency of seizure control (based on the underlying illness and the type and duration of seizure activity), the pharmacokinetics of various drugs (especially the rapidity of anticonvulsant effect), drugs already used and failed, and potential complications of treatment, especially hypotension and the risk for prolonged mechanical ventilation (algorithm 1). Clinicians should generally use medications they and the care team are familiar with in order to avoid unintended complications of therapy.

Concomitant antiseizure medications — In addition to drugs administered by continuous intravenous (IV) infusion, one or more longer-acting antiseizure medications should be administered in an effort to achieve and maintain seizure control and increase the likelihood of eventual tapering of the continuously infused drug.

It is critical that high therapeutic levels of at least one longer-acting antiseizure medication be maintained during infusion therapy and before tapering continuous infusions.

Antiseizure medications commonly used in this setting include levetiracetam, fosphenytoin/phenytoin, valproate, phenobarbital, and lacosamide. (See "Convulsive status epilepticus in adults: Management", section on 'Second therapy: Antiseizure medications'.)

Goals of infusion therapy — We generally titrate the infusion of the highly sedating antiseizure medication (midazolam, propofol, or pentobarbital) to aim for complete seizure control, both clinically and on the EEG. There is no conclusive evidence that a burst suppression pattern on EEG is necessary, while more suppression equates to more sedation and a longer intensive care unit course of treatment. Still, the EEG must be followed closely, as recurrent seizures often appear on the EEG before they are evident clinically.

The optimal electroclinical endpoint of treatment for RSE has not been studied rigorously, and it is uncertain whether the goal should be simple cessation of both clinical and electrographic seizures, or some degree of suppression of cerebral activity (eg, a burst suppression pattern on EEG). In one study of 35 patients on pentobarbital, three patients had EEGs showing simply freedom from seizures, and all survived [25]. Of the others, the 20 patients suppressed to the point of a "flat" background did substantially better than the 12 patients with a burst suppression pattern. However, this was a retrospective, nonrandomized study, and it is possible that the patients who were thought to have a better chance of survival were treated more aggressively. Other studies found that outcome of RSE was not associated with the depth of EEG suppression [26,27].

In a 2002 systematic review of drug therapy for RSE, patients treated with the goal of EEG background suppression (mostly with pentobarbital) had a lower likelihood of breakthrough seizures compared with those treated with the goal of clinical and electrographic seizure control (mostly with midazolam or propofol) (4 versus 53 percent) [28]. Correspondingly, patients treated to EEG background suppression had a greater likelihood of significant hypotension compared with those treated to seizure suppression (76 versus 29 percent). Of note, most such hypotension was readily managed with fluids and, occasionally, pressors. Mortality was high in both groups (48 percent) but did not appear to differ based on the electroclinical endpoint of therapy.

Efficacy and adverse effects — Whereas there is reasonable agreement about the initial treatment of generalized convulsive status epilepticus (GCSE), the optimal treatment of RSE is more controversial; there are no randomized trials comparing various treatments.

Serious potential complications of treatment for RSE in intubated patients with infusion of highly sedating medications include hypotension, prolonged mechanical ventilation, respiratory tract infections, propofol infusion syndrome, rhabdomyolysis, metabolic acidosis, and kidney failure [29].

A systematic review of drug therapy for RSE assessed data on 193 patients from 28 studies in an attempt to compare efficacy [28]. Pentobarbital treatment was associated with a lower frequency of breakthrough seizures (12 percent) compared with propofol or midazolam treatment (42 percent), but a higher frequency of hypotension, defined as a systolic blood pressure <100 mmHg (77 percent, versus 34 percent with propofol or midazolam). Overall mortality was 48 percent, but there was no association between drug selection and the risk of death. In a retrospective study of 107 patients, the outcome of refractory status was not associated with the choice of antiseizure therapy [26]. In a small randomized trial that was stopped early due to slow accrual, barbiturates and propofol were associated with similar rates of seizure control and hypotension, but patients receiving barbiturates had significantly longer mechanical ventilation times [30].

Of note, most pentobarbital trials were done before 1995, usually included only intermittent EEG recordings, and typically aimed for a burst suppression or more suppressed EEG background as the goal of treatment. Most midazolam and propofol studies published subsequently have used continuous EEG monitoring and aimed for seizure suppression rather than burst suppression on the EEG. Consequently, the available data are probably not comparing equally effective doses of these drugs. It is possible that pentobarbital would cause less hypotension if not titrated to such a suppressed EEG, while propofol and midazolam might lead to fewer relapses if used more aggressively, although possibly causing more hypotension.

Unlike most other drugs used in the treatment of refractory SE, ketamine does not suppress cardiovascular function and may have advantages for patients at risk of hypotension.

Specific treatments

Midazolam — Midazolam is a water-soluble, rapidly acting benzodiazepine that can control seizures within minutes [31,32]. It is generally initiated with a 0.2 mg/kg bolus given by IV infusion at a rate of 2 mg/min. Additional boluses should be given every five minutes until seizures stop (up to a maximum of 2 mg/kg), followed by a continuous infusion of 0.1 mg/kg/hour, which can be titrated upwards to as high as 3 mg/kg/hour.

If this is unsuccessful within 45 to 60 minutes, a propofol or pentobarbital infusion should be started. (See 'Propofol' below and 'Pentobarbital' below.)

Hypotension may be less common than with pentobarbital or thiopental [33] but commonly occurs at higher doses. The short half-life of midazolam (one to four hours) can increase markedly after days of use [34]. Tachyphylaxis is common, and the anticonvulsant effects of midazolam can cease rapidly when it is stopped. Withdrawal seizures and recurrent status epilepticus are therefore an important concern. Relapses of status epilepticus may be less frequent, and outcome may be better, when higher doses of midazolam are used [35].

Propofol — Propofol is a highly lipophilic phenol derivative and gamma-aminobutyric acid A (GABA-A) agonist with anticonvulsant properties. The drug is unrelated to any of the currently used barbiturate, opioid, benzodiazepine, or imidazole IV anesthetic agents. Hypotension and respiratory depression may complicate its use. (See "Sedative-analgesia in ventilated adults: Medication properties, dose regimens, and adverse effects", section on 'Propofol'.)

High-quality data for propofol dosing in the treatment of status epilepticus are limited; experience has been reported in several small studies [30,31,36,37]. As an example, one study compared the results of treatment with propofol (n = 8) or high-dose barbiturates (n = 8) in patients with RSE [38]. Termination of seizures was faster among successfully treated patients in the propofol group (mean 3 minutes versus 123 minutes with high-dose barbiturates).

Propofol infusion is started with a 1 to 2 mg/kg loading dose administered by IV infusion over approximately five minutes and repeated (0.5 to 2 mg/kg) until seizures stop, up to a maximum total dose of 10 mg/kg. After the initial loading dose, a continuous propofol infusion should be started at 20 mcg/kg per minute (1.2 mg/kg per hour) and titrated over the next 20 to 60 minutes to maintain a seizure-free state, though some experts aim for a burst suppression pattern on the EEG.

For breakthrough status epilepticus, a bolus of 0.5 to 2 mg/kg can be given every three to five minutes in addition to increasing the continuous infusion rate by 5 to 10 mcg/kg per minute (0.3 to 0.6 mg/kg per hour) every five minutes until seizures are controlled by clinical observation and EEG response. Infusion rates of up to 170 to 200 mcg/kg per minute (10 to 12 mg/kg per hour) may be required but should not be maintained for more than 48 hours because of the risk of the propofol infusion syndrome [38].

Adverse effects of propofol include dose-dependent hypotension, and the drug tends to be poorly tolerated in patients who are hypotensive, hypovolemic, or older. Therefore, for patients receiving propofol, it is important to correct hypovolemia with IV fluids, monitor blood pressure, and start vasopressors if needed to achieve effective blood pressure and avoid end-organ hypoperfusion.

The propofol infusion syndrome consists of rhabdomyolysis, severe metabolic acidosis, and cardiac and renal failure [39]. It appears to be more common with prolonged use (over 48 hours), in children, and with infusion rates of greater than 5 mg/kg per hour [40-42]. To avoid this complication, some recommend keeping the dose below 5 mg/kg per hour and for under 48 hours, especially in patients with acute neurologic insults such as head injury [39]. Others increase the dose to at least 15 mg/kg per hour as needed, at least for shorter periods [38]. The risk of acidosis may be increased in patients receiving carbonic anhydrase inhibitors such as acetazolamide, topiramate, or zonisamide. Arterial blood gases and serum levels of creatine phosphokinase (CPK), lactic acid, triglycerides, amylase, and lipase should be followed, and cardiovascular function monitored carefully during the continuous infusion. (See "Sedative-analgesia in ventilated adults: Medication properties, dose regimens, and adverse effects", section on 'Propofol-related infusion syndrome'.)

If seizures are controlled with propofol, the effective infusion rate should be maintained for 24 hours and then tapered at a rate of 5 percent per hour. This helps to prevent rebound seizures that commonly occur with abrupt propofol discontinuation. It is critical that high therapeutic levels of at least one longer-acting antiseizure medication are maintained prior to tapering the propofol in order to reduce the risk of seizure recurrence. (See 'Continuous EEG monitoring' above and 'Duration of continuous infusions' below.)

Treatment with propofol should generally be considered unsuccessful if it does not terminate seizure activity within approximately 60 minutes. In this case, switching to or adding a benzodiazepine drip (midazolam) or a high-dose barbiturate (ie, pentobarbital or thiopental) infusion may help. (See 'Pentobarbital' below.)

Pentobarbital — An initial dose of 5 mg/kg of pentobarbital should be infused over approximately 10 minutes (maximum rate 50 mg/minute). If seizure activity continues, additional 5 mg/kg re-bolus should be administered with careful attention to the EEG and hemodynamic status. This is followed by a continuous infusion of 1 mg/kg per hour, titrated as needed by 0.5 to 1 mg/kg per hour every 12 hours up to 5 mg/kg per hour to achieve seizure control or a burst suppression pattern on EEG.

Hypotension is common at this point, and many patients will need vasopressor support (typically phenylephrine or dopamine), as well as crystalloid infusions. Hypotension severe enough to necessitate stopping pentobarbital is uncommon. The mortality rate associated with barbiturate coma is high, although not necessarily higher than with equipotent doses of propofol or midazolam; contributing factors include the severity of the underlying illnesses causing RSE, adverse hemodynamic effects (hypotension, cardiac depression), a high incidence (~30 percent) of associated pneumonia, hepatoxicity, ileus, and immune dysfunction, possibly due to treatment [43]. The half-life of pentobarbital is 15 to 60 hours and can increase with prolonged use, so there is always prolonged sedation and an inability to assess the patient clinically.

If seizures are terminated with pentobarbital, the infusion is typically maintained for at least 24 hours before tapering. Many clinicians continue pentobarbital (or other continuous infusions) for longer when prolonged rapid and rhythmic epileptiform discharges suggest ongoing seizure activity, but continuing for frequent epileptiform discharges alone is not advisable.

Before tapering pentobarbital, high therapeutic concentrations of other antiseizure medications (eg, phenytoin, levetiracetam, valproate, phenobarbital, or others) should be maintained. Phenobarbital may be particularly useful in patients who develop recurrent seizures as pentobarbital is weaned [43]. Serum levels of pentobarbital are not particularly useful during weaning, but they can be useful after the medication has been discontinued to show if there is still a high level of pentobarbital accounting for a patient's unresponsiveness. (See 'Continuous EEG monitoring' above and 'Duration of continuous infusions' below.)

Ketamine — Ketamine, an N-methyl-D-aspartate (NMDA) antagonist that blocks the excitatory neurotransmitter glutamate, has promise as a treatment for RSE [44-49]. Glutamate antagonists might be particularly helpful in the later phases of status epilepticus when GABA agonists or promoters (eg, benzodiazepines and barbiturates) have lost some effectiveness and excessive glutamatergic activity may perpetuate seizures [50,51].

A typical loading dose of ketamine is 2 mg/kg, followed by an infusion of 1.5 to 10 mg/kg per hour, titrated to suppression of electrographic seizures. Optimal dosing has not yet been defined. Unlike most other drugs used in the treatment of refractory SE, ketamine does not suppress cardiovascular function and may have advantages for patients at risk of hypotension. (See "Procedural sedation in adults in the emergency department: Medication selection, dosing, and discharge criteria", section on 'Ketamine sedation'.)

Other antiseizure medications for RSE — Other antiseizure medications can be used for RSE when first-line medications fail.

●Topiramate – Topiramate is a broad-spectrum antiseizure medication. When administered via nasogastric tube in doses of up to 1600 mg/day, it appears to have some efficacy in RSE, as reported in retrospective case series [52-55]. Topiramate can cause a metabolic acidosis, which is of particular concern in patients also receiving propofol. It is not available in an IV form.

●Clobazam – Clobazam may be useful as adjunctive treatment for RSE when given enterally by nasogastric tube [56]. It has effects similar to those of other benzodiazepines, with a rapidity of onset that is intermediate between that of lorazepam and that of diazepam. Its duration of action is more prolonged than that of diazepam. (See "Convulsive status epilepticus in adults: Management", section on 'First therapy: Benzodiazepines'.)

●Brivaracetam – A retrospective study of 56 patients with status epilepticus found that treatment with brivaracetam (50 to 300 mg IV loading dose) was associated with seizure resolution in 57 percent; no severe adverse events were reported [57].

●Perampanel – Perampanel, an alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) receptor antagonist, is an oral drug and cannot be given intravenously, but its mechanism of action is different from other antiseizure medications used for status epilepticus, which may be an advantage for patients with RSE. In a retrospective study of 81 patients with refractory or super-refractory status epilepticus, perampanel treatment was associated with clinical improvement and resolution of seizures on EEG within 72 hours of starting treatment in 33 percent [58]. The median starting dose of perampanel in this study was 4 mg/day (range 2 to 36 mg/day), exceeding the recommended starting dose of 2 mg/day and the maximum maintenance dose of 12 mg/day used in the outpatient setting. Similarly, in a systematic review of 21 studies with 369 cases of status epilepticus, the proportion considered perampanel responders was 37 percent [59].

Immunomodulatory therapy — Although relatively rare, inflammatory or autoimmune etiologies such as anti-NMDA receptor encephalitis are important to consider in patients with new-onset refractory status (NORSE) with no etiology identified on the initial comprehensive evaluation, as early institution of immunomodulatory therapies (eg, glucocorticoids, intravenous immune globulin [IVIG]) may improve outcomes [60-63]. (See 'New-onset refractory status epilepticus' above.)

The proportion of cases of GCSE that are ultimately identified as having an autoimmune or paraneoplastic etiology ranges from 2 to 6 percent, with the higher estimate drawn from cases of RSE [61,62]. The diagnosis and treatment of paraneoplastic and autoimmune encephalitis are discussed in detail elsewhere. (See "Overview of paraneoplastic syndromes of the nervous system" and "Autoimmune (including paraneoplastic) encephalitis: Clinical features and diagnosis".)

NORSE often becomes super-refractory status epilepticus (see 'Super-refractory status epilepticus' above). Many cases of NORSE appear to have an immune, autoimmune, or inflammatory etiology or basis (table 1), and immunomodulatory therapy should generally be considered within the first 24 to 72 hours from the onset of status epilepticus [63]. First-line immune therapy such as IV glucocorticoids (eg, methylprednisolone 20 to 30 mg/kg [maximum 1 g] per day for three to five days), IVIG 2 g/kg given over two to five days, or plasmapheresis can be helpful [17].

If the response of NORSE or FIRES to initial immune therapies is incomplete, ketogenic dietary therapy should be started in children (and offered to adults) within one week from the onset of status epilepticus, and second-line immunosuppressive therapy (eg, rituximab, tocilizumab, or anakinra) should also be started [63-66]. The available evidence does not support the use of one of the secondary agents over the other, but expert opinion favors rituximab for most cases in which a pathogenic antibody is identified. (See 'New-onset refractory status epilepticus' above.)

Inhalational anesthetics — Inhalational anesthetics have been used for refractory GCSE [67], but are seldom used in the modern era given the profusion of so many other treatment options. Isoflurane and desflurane were reported to be effective in some highly refractory cases [68,69], but hypotension is a common problem with these agents, and relapses of seizures and status are relatively frequent upon medication withdrawal. They can provide definitive treatment but may be difficult to wean. Halothane has been avoided because of possible organ toxicity, and there is concern that enflurane can precipitate seizures.

Nonpharmacologic interventions — There are case reports and small case series describing a variety of other treatments for RSE, including vagus nerve stimulation [70,71], surgical approaches [72-74], transcranial magnetic stimulation [75], electroconvulsive therapy [10,76], and the ketogenic diet [77-79]. (See "Ketogenic dietary therapies for the treatment of epilepsy", section on 'Super-refractory status epilepticus'.)

Duration of continuous infusions — The duration of treatment is inadequately studied. In general, infusions are typically continued for 24 hours of clinical and electrographic seizure suppression and then gradually tapered over 12 to 24 hours. Although often tapered, pentobarbital has a very long half-life and does not necessarily need to be tapered. One retrospective study of patients on pentobarbital raised the possibility that a prolonged period of seizure and EEG suppression might be beneficial [25]. Another retrospective study of 182 patients with RSE who were treated mainly with propofol reported that higher doses but shorter duration of therapeutic coma were associated with better outcomes [80].

It is critical that high therapeutic levels of at least one longer-acting antiseizure medication be maintained before tapering continuous infusions. Insufficient use of longer-acting antiseizure medications likely increases the risk of relapse. Treatment with additional antiseizure medications started before and continuing at the time of tapering infusions is assumed in most trials, but this also has not been studied prospectively. In a study of 35 patients weaned from pentobarbital, all were maintained on phenytoin; those also on phenobarbital did substantially better when levels were above 15 microgram/mL [25].

Breakthrough seizures — When electrographic seizures reappear, they are generally not a benign finding, but rather predict a relapse of clinical seizures and status epilepticus [81,82].

●Change in treatment – Breakthrough seizures usually warrant an increase in treatment. It is common practice to re-treat with higher doses of the continuous infusional treatment or for longer at doses that were successful earlier, and then have additional antiseizure medications (or higher levels of earlier antiseizure medications) on board before the next attempt at tapering. Isolated epileptiform discharges, however, do not necessitate more treatment [25].

Nevertheless, prolonged intensive care unit stays are of great concern [83], and a few seizures (especially if relatively short and infrequent and without significant clinical manifestations) may have to be tolerated in order to wean the major sedating drugs.

●EEG interpretation – EEG interpretation can become difficult and controversial at this stage, and it is often hard to determine whether a given electrographic pattern is indicative of ongoing seizures and contributing to the patient's clinical deficit or more of an "interictal" pattern and not necessary to suppress. The involvement of an experienced epileptologist and electroencephalographer is crucial and may require monitoring the evolution of the EEG pattern over time. (See "Nonconvulsive status epilepticus: Classification, clinical features, and diagnosis", section on 'Uncertain EEG patterns in critical illness'.)

●Good recovery may require prolonged treatment – Treatment should not be abandoned prematurely for young patients in prolonged status epilepticus, or for any patient when no devastating cause has been found; some recover. There is no particular duration of status epilepticus or number of failed trials after which further treatment is uniformly futile, particularly in young patients and those without a severe underlying etiology. There are many reports of patients treated for RSE for weeks to months with good recovery [16,84-87].

In a randomized trial, induced hypothermia showed no evidence of benefit when added to standard therapies for the initial treatment of convulsive status epilepticus [88].

OUTCOMES — The outcome of RSE is often poor, with mortality rates ranging from 19 to 60 percent [4,5,84,89-91].

●Prognostic factors – As with status epilepticus in general, the most important prognostic factors are age, etiology, and medical comorbidities. The high mortality of patients with RSE is attributed mainly to the underlying etiology [28,31,82,92,93]. Seizure duration may also affect prognosis, although it is difficult to determine whether this is a factor independent of underlying etiology, particularly beyond the first few hours of status [8,89,94]. Prolonged seizures in patients with super-refractory status epilepticus have been associated with subsequent brain atrophy on serial magnetic resonance imaging scans [95].

In a single-center retrospective study that included 111 patients with RSE, the mean duration of RSE was 101 hours, and the overall in-hospital mortality rate was 38 percent [4]. In adjusted analysis of the entire cohort, hypoxic encephalopathy and brain tumors were independent risk factors for in-hospital death. Additional factors that were associated with poor outcome included longer seizure duration and presentation with nonconvulsive status epilepticus (NCSE) in coma. (See "Nonconvulsive status epilepticus: Classification, clinical features, and diagnosis", section on 'NCSE in coma'.)

Some case control studies have raised concern that the highly sedating (anesthetic) drugs themselves may also contribute to poor outcomes, independent of confounding factors such as age, etiology, and seizure severity [96,97]. Nevertheless, there are several reports of status epilepticus lasting over a month with reasonable recovery, depending mostly on etiology [16,82,84,85].

●Prognosis of NORSE – Patients with new-onset refractory status epilepticus (NORSE) often do not respond readily to antiseizure medications. Mortality and morbidity are high; patients with NORSE and other particularly refractory types of status epilepticus have fatality rates of 20 to 30 percent or more, and under 25 percent return to their prior baseline health [22,98]. However, many surviving patients have at least partial recovery with longer-term follow-up.

One multicenter retrospective study identified 130 cases meeting criteria for NORSE, with no etiology identified during the initial evaluation, who were admitted to 13 academic medical centers over a five-year period [62]. With additional testing beyond the first 48 hours, approximately one-half of patients had an etiology identified and half remained cryptogenic. The most common identified etiologies were autoimmune (19 percent) and paraneoplastic (18 percent). The mortality rate was 22 percent, and 62 percent of patients had a poor functional outcome at discharge (defined as a modified Rankin scale score >3). Among 63 surviving patients with post-discharge follow-up available (median nine months), over half had improved functional status, including 79 percent with good or fair outcome at last follow-up. Over 90 percent of surviving patients remained on antiseizure medications, including 37 percent with recurrent seizures.

A later retrospective study of 252 adults admitted to a neurologic intensive care unit for RSE identified 27 with NORSE [99]. In this group, 26 met criteria for new-onset, super-refractory status epilepticus (NOSRSE), defined by ongoing or recurrent seizures 24 hours or more after the onset of highly sedating (anesthetic) therapy [99]. Of the 26 patients with NOSRSE, 19 cases (73 percent) were cryptogenic; anti-N-methyl-D-aspartate (NMDA) receptor encephalitis was diagnosed in four patients (15 percent), infectious encephalitis in two (8 percent), and acute disseminated encephalomyelitis (ADEM) in one (4 percent). At discharge, only 6 patients (23 percent) had a good or fair outcome, while 14 (54 percent) had severe disability and 6 patients (23 percent) had died. Outcomes were particularly poor for patients with an identified etiology; the patients with infectious encephalitis and ADEM died, and the patients with anti-NMDA receptor encephalitis survived with severe disability.

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Seizures and epilepsy in adults".)

SUMMARY AND RECOMMENDATIONS

●Approach to treatment – The treatment of convulsive status epilepticus in adults is outlined in the algorithm (algorithm 1). For patients with generalized convulsive refractory status epilepticus (RSE), essential management steps include emergency consultation with neurology, transfer to an intensive care unit with continuous electroencephalography (cEEG) capability, rapid sequence intubation and mechanical ventilation, and highly sedating pharmacologic therapy to suppress seizures. For patients with focal RSE, a less aggressive approach is reasonable; treatment with conventional antiseizure medications should be exhausted before using highly sedating medications, particularly in nonintubated patients. (See 'Approach to treatment' above and 'Neurology consultation and intensive care' above.)

●Continuous EEG – cEEG monitoring is critical during the treatment of RSE in order to confirm that seizures have been treated adequately, to guide use of antiseizure medications and assess the level of suppression achieved, and to monitor for relapse of seizures as infusions are tapered. (See 'Continuous EEG monitoring' above.)

●Infusion therapy – For patients with RSE, we suggest initial therapy with midazolam (Grade 2C); reasonable alternatives include propofol and pentobarbital. In addition, one or more longer-acting antiseizure medications (eg, levetiracetam, fosphenytoin/phenytoin, valproate, phenobarbital, and/or lacosamide) are given in an effort to achieve and maintain seizure control and increase the likelihood of eventual tapering of the continuously infused drug. (See 'Choice of infusion agent' above and 'Concomitant antiseizure medications' above and 'Specific treatments' above.)

●Treatment goals – We generally titrate the infusion of the highly sedating antiseizure medication(s) to aim for complete seizure control both clinically and by cEEG; there is no conclusive evidence that a burst suppression pattern on EEG is necessary, and more suppression equates to more sedation and a longer intensive care unit course of treatment. Still, the EEG must be followed closely, as recurrent seizures often appear on EEG before they are evident clinically. (See 'Goals of infusion therapy' above.)

●Adverse effects – Potential complications of treatment for RSE with infusion of highly sedating medications include hypotension and prolonged mechanical ventilation. (See 'Efficacy and adverse effects' above.)

●Duration of infusions – The optimal duration of treatment for RSE is not well established. In general, infusions are typically continued for 24 hours of clinical and electrographic seizure suppression and then gradually tapered over 12 to 24 hours. (See 'Duration of continuous infusions' above.)

●Breakthrough seizures – Breakthrough seizures usually warrant an increase in treatment with higher doses of the continuous infusional treatment, or for longer at doses that were successful earlier, with additional antiseizure medications (or higher levels of earlier antiseizure medications) on board before the next attempt at tapering. (See 'Breakthrough seizures' above.)

●Outcomes – The outcome of RSE is often poor. The risk of mortality with RSE is mainly associated with the underlying etiology and is increased with hypoxic encephalopathy and brain tumors.

New-onset refractory status epilepticus (NORSE) with no etiology identified during the initial evaluation (table 1) is also associated with high morbidity and mortality, particularly for patients ultimately determined to have an underlying cause such as infectious or autoimmune encephalitis.

Yet treatment should not be abandoned in patients with prolonged status epilepticus deemed to have potential for recovery if the RSE can eventually be controlled (eg, young patients without devastating brain injury). Meaningful functional recovery is sometimes possible even after weeks of treatment with highly sedating drug infusions. (See 'Outcomes' above.)