Management of Status Epilepticus in Adults

From the Johns Hopkins Hospital, Baltimore, MD (Dr. Ramadan), and the Henry Ford Hospital, Detroit, MI (Dr. Varelas).

Abstract

• Objective: To review the management of status epilepticus (SE).
• Methods: Review of the literature.
• Results: SE is a relatively common condition that accounts for 3% to 5% of all emergency department evaluations for seizure disorders and occurs in 2% to 16% of all epilepsy patients. The 3 most common etiologies are low levels of antiepileptic drugs, remote symptomatic etiologies, and cerebrovascular accidents. The majority of SEs are convulsant, but there is growing awareness of non-convulsive SEs, which can be diagnosed only via electroencephalogram. Management, which must be initiated at the earliest possible time, has evolved to incorporate pre-hospital measures and 4 treatment stages, with supportive measures and benzodiazepine administration remaining the mainstay initially and followed by older and newer antiepileptic drugs and anesthetics for resistant cases.
• Conclusion: SE is a neurological emergency that still carries significant mortality and morbidity if not treated immediately and properly.

Key words: status epilepticus; seizures; convulsive status epilepticus; nonconvulsive status epilepticus.

Status epilepticus (SE) is a relatively common condition that accounts for 3% to 5% of all emergency department (ED) evaluations for seizure disorders and occurs in 2% to 16% of all epilepsy patients [1]. It remains a major neurological emergency that, if not properly and timely treated, leads to death or permanent neurological injury. Since most of patients with convulsive SE are admitted to the hospital via the ED and are then transferred to the intensive care unit (ICU), our focus in this review will be on the latter.

Although only a handful prospective, randomized studies have been reported, guidelines on SE have been published in Europe [2] and the US [3,4]. In this paper, we review the evolving definition and types of SE, its incidence, etiology, and pathophysiology, its diagnosis and treatment algorithms, and its outcome. Our goal is to provide the reader with a concise but thorough review of this still lethal neurological emergency.

Definitions

The International Classification of Epileptic Seizures had previously defined SE as any seizure lasting ≥ 30 minutes or intermittent seizures lasting for > 30 min without recovery of consciousness interictally [5,6]. More recently, a duration of 5 or more minutes of (a) continuous seizures or (b) 2 or more discrete seizures with incomplete recovery of consciousness in-between, proposed by Lowenstein [3,7], offers the advantage of incorporating new knowledge. The shortening of the convulsive period to 5 minutes was based on the fact that the majority of tonic-clonic seizures last for only 1 to 2 minutes, that those lasting > 5 minutes do not stop spontaneously [8], that permanent neuronal injury occurs before 30 minutes, and that refractoriness to treatment increases with longer seizure duration [9].

Refractory SE (RSE) has been defined as SE not controlled after adequate doses of an initial benzodiazepine followed by a second acceptable antiepileptic drug (AED) or SE not controlled after the initial parenteral therapy with a minimum number of standard “front-line” AEDs (either 2 or 3) or SE with a minimum duration of seizures that persist despite treatment (eg, at least or 2 hours) [3,10]. Super-refractory SE (SRSE) is defined as SE that continues or recurs 24 hours or more after the onset of anesthetic therapy or recurs on the reduction or withdrawal of anesthesia [11].

Non-convulsive SE (NCSE) is defined as the presence of altered consciousness or behavior for ≥ 30 minutes, the absence of overt clinical signs of convulsive activity during that period, and the electroencephalographic (EEG) confirmation of seizures or activity that responds to treatment together with improvement of consciousness [12–15]. Two major types of NCSE can be encountered: the one in patients with epileptic encephalopathy/coma and the one in patients with absence or complex partial seizures, who are not usually admitted to ICU and are functional yet impaired. Because of the confusion between these 2 extremes in the NCSE spectrum, working criteria for standardization of reporting, utilizing the frequency of electroencephalographic epileptiform discharges or delta/theta waveforms have been proposed [15]. A recent compendium of 123 cases of NCSE with clinical descriptions and EEG patterns following a syndromic classification approach has also been published [16].

Types of SE

Three major categories of SE have been described: generalized convulsive SE (GCSE), focal motor SE (FMSE or epilepsia partialis continua [EPC]) of Kojevnikov, and NCSE. GCSE and FMSE are easily recognized due to overt convulsions. NCSE, however, has a more obscure phenotype and can be subdivided into a spectrum encompassing typical absence and complex partial SE, atypical absence SE and tonic SE (usually in children with learning disabilities), epileptic behavioral disturbance and psychosis, including Balint–like syndrome [17], confusional states or delirium with epileptiform discharges) and SE in coma (after significant brain injuries, such as hypoxia-ischemia, most commonly encountered in ICUs) [13,18]. The 2 extremes in this NCSE spectrum have completely different prognoses, with absence SE the most benign and SE in coma the most dismal.

Lastly, SE presents either spontaneously or can be “semi-intentional” iatrogenic, encountered either in the neuro-ICU or epilepsy monitoring unit, when AEDs are withdrawn under continuous EEG recording in order for seizures to emerge and be recorded with surface or intracranial electrodes.

Incidence of SE

In a prospective population-based epidemiological study, the incidence of SE was estimated at 41–61/100,000 patients/year. For the US, this translates to 125,000 to 195,000 episodes per year [19].

The highest incidence of SE occurs during the first year of life and during the decades beyond 60 years, and is also dependent on the SE subtype. Partial SE occurs in 25% of cases of SE and NCSE accounts for another 4% to 26 % [19,20], but the incidence for the latter is considered an underestimate due to the need for continuous EEG monitoring (which is not widely available). For example, NCSE was discovered in no patient with acute stroke [21], 8% of comatose ICU patients [22], 7% of patients with intracerebral hemorrhage [23], 3% to 8% of patients with subarachnoid hemorrhage [24–26], 6% of patients with metastatic cancer [27], and 6% of patients with head trauma [28].

The incidence of RSE and SRSE is also unknown. In a recent retrospective study from a neuro-ICU in a West China hospital, the percentage of non-refractory SE, RSE, and SRSE were 67.3%, 20.4% and 12.2%, respectively [29]. Other retrospective studies have shown that 12% to 43% of SE cases become refractory [30–33] and that approximately 10% to 15% of all cases of hospital-admitted SE will become super-refractory at some point, but no prospective studies have been published.

Risk factors that have been identified for RSE are encephalitis as a cause, severe consciousness impairment, de novo episodes of SE, delay in initiation of treatment, NCSE, and focal motor seizures at onset [30,32,34,35]. In a more recent study from ICU patients in Switzerland and the US, acute SE etiology (traumatic brain injuries, cerebrovascular accidents, meningoencephalitis, brain tumors, surgical brain lesions, exposure to, or withdrawal from, recreational drugs, prescription drugs, alcohol, metabolic disturbances and fever), coma/stupor, and serum albumin < 35 g/L at SE onset were independent predictors for RSE [36].

Etiology of SE

The 3 most common etiologies for SE are low levels of antiepileptic drugs (AEDs) in 34% of the cases (usually due to noncompliance), remote symptomatic etiologies (history of neurological insults remote to the first unprovoked SE episode, 24%), and cerebrovascular accidents (ischemic and hemorrhagic strokes, 22%). These are followed by hypoxia (13%) and metabolic disturbances (15%). Because 82% of patients in the remote group have a history of cerebrovascular disease, almost 50% have either acute or remote cerebrovascular disease as etiology of SE [19].

In general ICUs, metabolic abnormalities can account for 33% of seizures, drug withdrawal for 33%, drug toxicity for 14.5%, and stroke for 9% to 39% [37,38]. In ICUs, sepsis remains a common etiology of electrographic seizures or periodic epileptiform discharges [39,40], and legal or illegal drugs, such as ciprofloxacin, levofloxacin, piperacillin/tazobactam, cefepime and carbapenems [41–43], lithium or theophylline intoxication, vigabatrin, tiagabine or crack/cocaine, are another [18] (especially when their metabolism is altered due to interactions with other drugs or when their excretion is impaired due to hepatic or renal failure).

Beyond these common causes of SE, a workup for rare etiologies should be entertained. In a systematic review of 513 papers on SE, 181 uncommon causes of SE were identified and subdivided into immunologically mediated disorders, mitochondrial diseases, rare infectious disorders, genetic disorders, and drugs or toxins [18,44].

The most recent knowledge in this category is the contribution of paraneoplastic or autoimmune conditions to a large percentage of previously cryptogenic pharmaco-resistant seizures or super-refractory SE, most in the context of limbic encephalitis. Many of these patients have never experienced seizures or SE before and a new acronym has been devised for them: new-onset refractory status epilepticus (NORSE), ie, a state of persistent seizures with no identifiable etiology in patients without preexisting epilepsy that lasts longer than 24 hour despite optimal therapy [45]. A growing array of autoantibodies against intracellular and surface or synaptic neuronal targets has been described in addition to the previous literature of Rassmussen’s encephalitis and Hashimoto’s encephalopathy [46]. The most common autoantibodies associated with seizures and SE include anti-Hu, anti-Ma2, anti-CV2/CRMP5, anti-Ri, ANNA3, anti-amphiphysin, anti-NMDA receptor, anti-LGI1 and CASPR2, anti-GABA-beta, anti-GluR3, anti-mGluR5 and alpha 3 ganglionic acetylcholine receptor [47,48]. The diagnosis frequently remains elusive due to lack of knowledge or absence of widespread availability of serologic testing (with sometimes weeks-long delay for the results to be available), but the response to treatment with removal of tumor, plasmapheresis, or immunomodulation and immunosuppression is often dramatic.