Status epilepticus (SE) treatment strategies vary substantially from one institution to another due to the lack of data to support one treatment over another. To provide guidance for the acute treatment of SE in critically ill patients, the Neurocritical Care Society organized a writing committee to evaluate the literature and develop an evidence-based and expert consensus practice guideline. Literature searches were conducted using PubMed and studies meeting the criteria established by the writing committee were evaluated. Recommendations were developed based on the literature using standardized assessment methods from the American Heart Association and Grading of Recommendations Assessment, Development, and Evaluation systems, as well as expert opinion when sufficient data were lacking.

CONTEXT: The optimal pharmacologic treatment for early convulsive status epilepticus is unclear. OBJECTIVE: To analyze efficacy, tolerability and safety data for anticonvulsant treatment of children and adults with convulsive status epilepticus and use this analysis to develop an evidence-based treatment algorithm. DATA SOURCES: Structured literature review using MEDLINE, Embase, Current Contents, and Cochrane library supplemented with article reference lists. STUDY SELECTION: Randomized controlled trials of anticonvulsant treatment for seizures lasting longer than 5 minutes. DATA EXTRACTION: Individual studies were rated using predefined criteria and these results were used to form recommendations, conclusions, and an evidence-based treatment algorithm. RESULTS: A total of 38 randomized controlled trials were identified, rated and contributed to the assessment. Only four trials were considered to have class I evidence of efficacy. Two studies were rated as class II and the remaining 32 were judged to have class III evidence. In adults with convulsive status epilepticus, intramuscular midazolam, intravenous lorazepam, intravenous diazepam and intravenous phenobarbital are established as efficacious as initial therapy (Level A). Intramuscular midazolam has superior effectiveness compared to intravenous lorazepam in adults with convulsive status epilepticus without established intravenous access (Level A). In children, intravenous lorazepam and intravenous diazepam are established as efficacious at stopping seizures lasting at least 5 minutes (Level A) while rectal diazepam, intramuscular midazolam, intranasal midazolam, and buccal midazolam are probably effective (Level B). No significant difference in effectiveness has been demonstrated between intravenous lorazepam and intravenous diazepam in adults or children with convulsive status epilepticus (Level A). Respiratory and cardiac symptoms are the most commonly encountered treatment-emergent adverse events associated with intravenous anticonvulsant drug administration in adults with convulsive status epilepticus (Level A). The rate of respiratory depression in patients with convulsive status epilepticus treated with benzodiazepines is lower than in patients with convul-

Benzodiazepines are safe and effective when administered by paramedics for out-of-hospital status epilepticus in adults. Lorazepam is likely to be a better therapy than diazepam.

No abstract

We retrospectively reviewed the clinical course of adult patients treated for generalized status epilepticus (SE) at the San Francisco General Hospital (SFGH) from 1980 to 1989. The review was designed to determine whether the etiologies of SE at our hospital have changed over the last two decades, and to investigate the relationships between etiology, response to anticonvulsant therapy, and short-term clinical outcome. Of 154 patients reviewed, the four leading etiologies for SE were anticonvulsant drug withdrawal (39), alcohol-related (39), drug toxicity (14), and CNS infection (12). This pattern was essentially unchanged from observations made at SFGH in the 1970s. Sixty percent of all patients responded to first-line drug treatment (usually phenytoin +/- diazepam), and the remainder required an additional agent (usually phenobarbital) for control of SE. The best response to anticonvulsants occurred in patients with SE related to tumor, anticonvulsant drug withdrawal, or refractory epilepsy, and the poor responders had anoxia, drug toxicity, CNS infection, or other metabolic abnormalities. Seventy-six percent of the patients had good outcomes. Of the 22 patients who died, SE was a likely cause of death in only two (ie, 1.3% of the entire study group). Metabolic abnormalities, stroke, and anoxia were associated with particularly poor outcomes compared with other etiologies. These observations show that the etiologies of SE have remained similar over two successive decades, and that the etiology of SE may help predict both the initial response to drug therapy and the short-term outcome.