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, 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.


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.


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.

Adapted from: Drislane FW. Status epilepticus. In: Schachter SC, Schomer DL, eds. The comprehensive evaluation and treatment of epilepsy. San Diego, CA: Academic Press; 1997. p. 149-172.
With permission from Elsevier (www.elsevier.com)
Authored By: 
Frank W. Drislane MD
Reviewed By: 
Thaddeus Walczak
Thursday, January 1, 2004