In generalized convulsive status epilepticus, early escalation to anesthetic drugs is justified because rapid seizure control is imperative to avoid the development of pharmacoresistance, neuronal injury, and systemic complications. Commonly used anesthetic ASDs are listed in Table 17.2. There is insufficient evidence to recommend one anesthetic ASD over another [20, 21]. There are three conventional choices—barbiturates (thiopental or its main metabolite, pentobarbital), midazolam, and propofol—although ketamine has become an alternative choice as experience with it has increased. One randomized controlled trial was attempted comparing thiopental and midazolam, but the trial was powered for 150 patients and recruited only 24 [22]. A systematic review of published (primarily uncontrolled) case series reported control of RSE without breakthrough seizures to be 42, 66, and 60%, respectively, for midazolam, propofol, and barbiturates [6].

As there are no randomized or controlled comparative data upon which to differentiate these choices, selection is based primarily on the advantages and adverse effect profile of each drug in relation to the comorbidities of the patient. It should be stated that all anesthetic ASDs are associated with high rates of infection [16, 17]. If an anesthetic ASD is initiated and titrated to typically adequate doses without achieving electrographic seizure control, an alternative anesthetic drug is usually added or substituted. According to recently published data from the global audit of treatment of refractory SE, the most widely used initial anesthetic ASD is midazolam (59%), followed by propofol (32%), and barbiturates (8%) [12].

Midazolam. Midazolam is a benzodiazepine administered via IV infusion which acts by binding to and enhancing the action of the GABA_A receptor. Onset of action occurs within minutes and it is relatively short-acting in non-obese patients with normal renal function (elimination half-life of 1.8–6.4 h). These properties make it ideally suited to prolonged use without accumulation, but accumulation may occur in adipose tissue and with renal insufficiency. Tachyphylaxis may develop, sometimes after only one day of use, necessitating gradually increasing doses to maintain seizure control. The propensity for breakthrough seizures to develop during treatment with midazolam has been shown in multiple studies [34, 35]. As midazolam is a strong respiratory depressant, mechanical ventilation is required, and hypotension requiring pressors occurs in 30–50% of patients [6, 29, 34]. In a systematic review of 28 studies describing 193 patients with RSE, 54 of whom were treated with midazolam, seizures recurred acutely after the loading dose in 20% of cases. Breakthrough seizures occurred after the first six hours of treatment in 51%, and withdrawal seizures occurred during weaning of midazolam in 63% [6]. More recently, a study compared 100 patients treated with a high-dose continuous midazolam infusion (median maximum dose 0.4 mg/kg/h, interquartile range (IQR) 0.2–1.0) to 29 historical controls at the same center treated with a lower dose midazolam protocol (median maximum dose 0.2 mg/kg/h, IQR 0.1–0.3) [29]. Withdrawal seizures, occurring within 48 h of drug discontinuation, were less frequent in the high-dose group (15 vs. 64%; odds ratio (OR) 0.10; 95% CI 0.03–0.27) and mortality was lower (40 vs. 62%; OR 0.34; 95% CI 0.13–0.92) compared with those in the low-dose group, despite a higher incidence of hypotension, and similar baseline patient characteristics and duration of midazolam infusion. The results of this study suggest that high doses of midazolam are safe and associated with fewer withdrawal seizures. The implications of the lower mortality are unclear given the historical controls and inability to account for other practice changes.

Propofol. Propofol is an anesthetic with ill-defined anti-seizure properties, which is thought to act by modulation of the GABA_A receptor, and possibly N-methyl-D-aspartate (NMDA) antagonism, at least in vitro [36]. Like midazolam, propofol is very short-acting and has a rapid onset of action. Other advantages include its intracranial pressure and cerebral metabolism lowering properties [37]. Pressors are required for treatment of hypotension in 22–55% of patients [6, 22, 38]. Apnea occurs in 50–84% of patients, and mechanical ventilation is required [39]. The most feared complication of propofol is the propofol infusion syndrome (PRIS), a syndrome of metabolic acidosis, rhabdomyolysis, renal failure, hyperkalemia, hypertriglyceridemia, and rapid cardiovascular collapse which results from a toxic effect on mitochondrial and cellular metabolic function. The incidence of PRIS is unknown, and estimates vary widely with the dose and duration of use [38, 40]. Risk factors include young age, high fat and low carbohydrate intake, concomitant catecholamine infusion or corticosteroid use, and prolonged high-dose infusions (≥80 mcg/kg/min, for ≥3 days) [40, 41]. In a study of 31 patients with RSE treated with propofol for a median 67 (range 2–391) hours with median cumulative doses of 12,850 (range 336–57,545) mg, three sudden cardiorespiratory arrests occurred without clear explanation. Two patients died and 11 additional patients exhibited features of PRIS despite careful monitoring for metabolic and cardiac changes [40]. It is therefore likely that the only way to avoid this potentially lethal complication is with the use of a protocol limiting its use to no more than 2 or 3 days at doses not higher than 80 mcg/kg/min. Treatment of PRIS is primarily supportive and includes stopping propofol, supporting the cardiopulmonary and renal systems, sometimes with cardiac pacing, renal replacement therapy, and extracorporal membrane oxygenation [42].

While significant clinical experience exists with propofol, data about efficacy is limited. One study examined the use of propofol in 27 consecutive episodes of RSE retrospectively, and found that breakthrough seizures occurred in 9/27 (33%) episodes, but in only two cases were the seizures severe enough to prompt substitution of an alternative anesthetic ASD [38]. In a systematic review of 28 studies describing 193 patients with RSE, 33 of whom were treated with propofol, seizures recurred acutely after the loading dose in 27% of cases. Breakthrough seizures occurred after the first six hours of treatment in 15%, and withdrawal seizures occurred during weaning in 46% of cases [6].

Barbiturates. Thiopental and its metabolite, pentobarbital, are barbiturate anesthetic drugs with strong anti-epileptic action. Their primary mechanism of action is to enhance transmission at the GABA_A receptor but they also lower the core body temperature and may have neuroprotective effects. The barbiturates have a strong sedative effect and are respiratory depressants, necessitating mechanical ventilation. At high doses, they can result in loss of all brainstem reflexes and an isoelectric EEG, mimicking brain death [43].

Barbiturates are virtually always effective in achieving initial seizure control. Nevertheless, because of their prolonged duration of action, it is this author’s opinion that they are not an ideal choice for first-line anesthetic therapy. There is a subset of patients with RSE who require only 24 h of anesthetic therapy and, upon correction of the etiology, are easily weaned from anesthesia, extubated, and discharged from the ICU within a 12–24 h period. If thiopental or pentobarbital is chosen as the first-line anesthetic therapy, the likelihood of the patient awakening and liberation from mechanical ventilation within that time frame is significantly reduced. This is due to their zero order kinetics, rapid redistribution, and resultant accumulation leading to a long half-life, prolonged recovery times [44], and longer duration of mechanical ventilation [22]. The barbiturates are metabolized by the liver, undergo autoinduction, and have many drug–drug interactions. Hypotension requiring pressors occurs in 29–77% [6, 45] of patients [22]. While less common, several other potentially serious systemic complications are specific to barbiturate. A relatively common complication of barbiturate infusions is adynamic ileus (Fig. 17.2), reported in 10% of patients [45]. When severe, bowel ischemia, and even perforation can result [22, 46]. Rarely, lingual edema (Fig. 17.3) can develop risking airway obstruction [47]. This gradually resolves after discontinuation of the drug. In <1% of patients, propylene glycol toxicity may develop which manifests as a progressive acidosis that resolves after drug discontinuation [45]. Rarely, pancreatic, gastric, or hepatic injury may develop due to systemic and splanchnic hypoperfusion, complications which more commonly occur in elderly patients [48].

In a retrospective review of 31 patients with SRSE treated with pentobarbital infusions, seizure control was achieved in 90% of patients but recurred in 48% upon weaning of the drug [45]. A systematic review of 28 studies describing 193 patients with RSE, 106 of whom were treated with pentobarbital, seizures recurred acutely after the loading dose in only 8% of cases. Breakthrough seizures occurred after the first six hours of treatment in 12%, and withdrawal seizures occurred during weaning in 43% of cases [6].

Ketamine. Ketamine is an NMDA antagonist—a potential advantage over the other anesthetic ASDs, as prolonged seizures are accompanied by pharmacoresistance to GABA agonists [49] but not to NMDA antagonists [50]. An additional advantage is its lack of respiratory depressant effects. Onset of action occurs within seconds, and it is relatively short-acting (elimination half-life of 2–3 h). Metabolism is hepatic and excretion is largely renal. Efficacy has been demonstrated in animal models, even in late stages of RSE [50, 51]. Experience in humans with RSE has been increasing in recent years. The largest published series is a multicenter retrospective review of 46 adults and 12 children totaling 60 episodes of RSE treated with ketamine [52]. In this series, ketamine was thought to have contributed to permanent control of RSE in 32% of cases, and transient control in an additional 13%, similar to the reported efficacy of the other anesthetic ASDs [6]. Interestingly, response rate was highest when ketamine was introduced early (as a third or fourth line agent). Still, an assessment of efficacy in a retrospective fashion, without controlling for the effects of other ASDs, treatment of the cause of RSE, and other factors, is questionable. The true value of this study lies in its confirmation of relative safety at the reported doses. Infusions of up to 10 mg/kg/h for up to 27 days were not associated with increased complications or mortality compared to patients receiving lower doses for fewer days [52]. Two patients in this series developed supraventricular tachycardia that resolved after drug discontinuation. One developed atrial fibrillation requiring amiodarone, and there was one incident of severe acidosis during coadministration of both high-dose midazolam and ketamine leading to discontinuation of the drug. Despite a call for earlier use [53], it is generally reserved for the most severe cases, usually after more than one anesthetic ASD has failed [50], a practice which is in line with current guidelines [53].

Once an anesthetic ASD has been initiated, the primary treatment goals are clinical and electrographic seizure suppression, and reversal of the cause of seizures. It is a common practice to titrate anesthetic ASDs to a predetermined EEG endpoint. Endpoints are controversial, and available evidence is conflicting [8, 10, 29, 54]. Options include complete background suppression (sometimes referred to as ‘isoelectric’ or ‘flat’), burst suppression, or seizure suppression. Determining how much to suppress requires a clinical judgement that balances the risks of increased suppression (very high doses of anesthetic ASDs are sometimes required to achieve a burst suppression or isoelectric EEG background, risking increased hypotension, and other systemic complications), with the benefit of increased seizure suppression. Continuous EEG monitoring has shown that seizures may still emerge from a burst suppression pattern, so it follows that greater suppression should confer better seizure control [55].

Once achieved, it is standard practice to maintain the desired EEG endpoint for 24–48 h prior to a slow withdrawal of anesthetic ASDs [53]. Prior to attempting the first anesthetic wean, 2–3 non-anesthetic ASDs (usually including the drug selected as second line therapy) should be initiated at high doses and titrated to achieve therapeutic levels. In patients at risk for development of adynamic ileus (i.e., patients receiving opiates or barbiturates), the IV route of administration is preferred to ensure reliable absorption. In other patients, ASDs may be administered enterally via a nasogastric or orogastric tube. No evidence exists to guide optimal ASD combinations in this setting. General considerations for drug selection include seizure type, systemic comorbidities, drug–drug interaction profiles, and avoidance of polypharmacy (>3 ASDs may add morbidity by increasing the risk of adverse effects without evidence of benefit).

In addition, the clinician must ensure that the underlying etiology has been addressed. By this time patients will have undergone at a minimum, a thorough history, noncontrast head computed tomography (CT) scan, comprehensive laboratory evaluation, and lumbar puncture for cerebrospinal fluid (CSF) analysis. If an etiology is identified, attempts are made to correct the etiology (e.g., reverse hypoglycemia), or at least initiate appropriate treatment when the etiology is not expected to resolve rapidly (e.g., fulminant bacterial meningitis). If an etiology has not yet been identified at this stage, this is the time to begin the search for more unusual causes of SE (e.g., complete autoimmune encephalopathy panels including NMDA receptor, and voltage gated potassium channel antibodies) and to consider initiation of empiric immunotherapy. (See also Chap. 8, “Unusual Causes of Status Epilepticus.”)

There is no evidence to guide the weaning of an anesthetic ASD. This author’s practice is to wean anesthetic drugs one at a time by 10% per hour. For example, if the desired EEG endpoint is achieved with midazolam 20 mg/h and propofol 50 mcg/kg/min, and maintained for 24–48 h, the midazolam dose would be decreased by 2 mg/h each hour until off, while carefully monitoring the EEG for seizure recurrence. When midazolam is successfully discontinued, the propofol wean would begin, decreasing the dose by 5 mcg/kg/min every hour until off. Deciding which anesthetic drug to wean first is somewhat arbitrary but clinical circumstances may dictate which drug to wean first. Patients exposed to prolonged barbiturate infusions are at higher risk for withdrawal seizures; this may be avoided by utilizing phenobarbital as one of the 2 or 3 non-anesthetic ASDs [56].

If seizures recur, usual practice is to resume anesthesia and reestablish EEG suppression. Occasionally breakthrough seizures that occur upon weaning of anesthesia will subside spontaneously. How long to observe, if at all, remains a clinical judgment. It is probably reasonable to observe and allow some electrographic seizures during an anesthetic wean, but if the frequency of breakthrough seizures does not decline gradually over time, anesthesia should be resumed. The same principle applies to other patterns that do not meet criteria for electrographic seizures, but are on the ictal–interictal continuum. As long as the EEG background continues to improve, and seizures are infrequent and declining in frequency, this author’s practice is to continue weaning. During this time, bolus doses of benzodiazepines and further optimization of the patient’s non-anesthetic ASD regimen may increase the likelihood of successful weaning.

If seizures continue or recur 24 h or more after the initiation of anesthetic therapy, the patient is considered to have reached the stage of ‘super-refractory status epilepticus.’ The majority of recommendations to guide treatment of this stage come from expert consensus. Issues that remain unresolved include what anesthetic drug to choose after failure of a weaning attempt and when to attempt weaning again. Barbiturates are often used as second line anesthetic ASDs [12]. They are reasonably well suited for long term use and are not prone to tachyphylaxis. Over time, the time between weaning attempts is increased, and after failure of several attempted weans, anesthesia is often continued for 5–7 days between weaning attempts.

Clinical features supportive of immune mediated SE include: (1) a well-defined clinical syndrome (e.g., limbic encephalitis or faciobrachial dystonic seizures), (2) subacute onset (maximal seizure frequency <3 months) of cryptogenic epilepsy, (3) cryptogenic RSE or new onset refractory status epilepticus (NORSE), (4) a viral prodrome, (5) antecedent psychiatric symptoms, (6) history of systemic autoimmunity, or (7) history of neoplasia. Supportive paraclinical features include: (1) evidence of central nervous system inflammation (e.g., CSF pleocytosis, elevated CSF protein, CSF oligoclonal bands, elevated CSF IgG index or synthesis rate, mesial temporal or parenchymal T2-weighted or fluid-attenuated inversion recovery sequence hyperintensities, or hypermetabolism on functional imaging), (2) extreme delta brush pattern on EEG, or serologic markers of systemic autoimmunity (e.g., antinuclear antibody or thyroid peroxidase antibody positivity) [58]. These patients should undergo comprehensive evaluation for neural-specific autoantibodies in the serum and CSF.

Multiple arguments in favor of early empiric initiation of immunotherapy in SRSE can be made. First, earlier initiation of immunotherapy confers a better outcome in autoimmune central nervous system diseases when compared with delayed initiation of therapy [59–61]. Second, autoimmune and paraneoplastic syndromes are the most common cause of cryptogenic RSE [62], also known as NORSE [63]. Finally, there is increasing evidence that inflammation plays an important role in epileptogenesis and activation of specific inflammatory signaling pathways (e.g., interleukin-1 receptor/toll-like receptor (IL-1R/TLR) pathway) [64–67].