Refractory status epilepticus (RSE) is defined as an ongoing seizure activity despite treatment with a first-line agent (benzodiazepines) and a second-line AED (phenytoin, valproic acid, levetiracetam, or phenobarbital) [12]. RSE has also been defined as “status epilepticus requiring general anesthesia” [13]. RSE can be diagnosed clinically if obvious clinical seizure activity is evident. If, however, this is not the case and the patient has altered awareness, electrographic confirmation becomes necessary.

Super refractory status epilepticus (SRSE) is defined as ongoing electrographic seizure activity despite treatment with first-, second-, and third-line AEDs. Clinical activity, if present, is subtle. SRSE has also been defined as “status epilepticus that continues 24 h or more after the onset of anesthesia, including those cases in which the status epilepticus recurs on the reduction or withdrawal of anesthesia” [13].

NCS and NCSE are often discussed together without any differentiation in their treatment paradigms. It is impossible to separate these two clinically, given that the diagnosis is based primarily on EEG findings. One of the principal components of the SE definition is the lack of complete clinical recovery to baseline in between seizures. This criterion cannot be used to differentiate NCS from NCSE due to the lack of clinical features except depressed mental status. Continuous EEG (cEEG) monitoring, however, can be used to differentiate these two conditions. In NCSE electrographic seizure activity is continuous or near continuous, while in NCS, there are discrete electrographic seizures separated slowing and interictal activity noted on EEG (Fig. 1). The distinction between NCS and NCSE is important as it may have implications on treatment.

The aggressiveness of treatment depends on the degree of neuronal injury that is ongoing. It is well understood that GCSE causes neuronal injury and should be aggressively treated. Conversely, single seizures are unlikely to cause brain damage and generally do not require an immediate change in treatment. NCS and NCSE fall in between these two extremes, and the degree of aggressiveness in treating these conditions is unclear (Fig. 2).

Several animal studies have led to further understanding of mechanisms underlying neuronal injury in SE although much of the pathophysiology remains poorly understood. The basic principle underlying the development of SE involves failure of endogenous mechanisms to terminate a seizure. This can be due to excessive excitation or from loss of endogenous inhibitory mechanisms. Injury during SE is postulated to occur in several proposed stages. In the initial milliseconds to seconds of seizure onset, neurotransmitter release, ion channel opening and closing, and protein phosphorylation may lead to a prolonged seizure. This is followed in the next seconds to min by alteration in receptor trafficking where there is a decrease in inhibitory gamma-aminobutyric acid (GABA) subunits and an increase in excitatory N-methyl-D-aspartate (NMDA) and α-amino-3-hydroxy-5-methylisoxazole-4-proprionic acid (AMPA) receptors. In the subsequent min to hours, there are changes in neuropeptide expression leading to increase in excitatory substance P and decrease in inhibitory neuropeptide Y. Days to weeks of continued seizure activity leads to genetic and epigenetic changes causing further neuronal damage [14]. Injury at neuronal level occurs concomitantly with changes in brain metabolism and systemic pathology. Initially, the sympathetic overdrive leads to a compensatory phase where cardiac output, blood pressure, and blood glucose increase. Continued seizure activity > 30 min leads to systemic decompensation with decrease in cerebral blood flow, glucose, and oxygen [1].

Although most of the animal models included GCSE, there is some evidence of neuronal injury in NCSE as well. In a series of experiments done in baboons where convulsive seizures were induced, only partial protection against neuronal injury was observed after paralysis of baboons, implying continued neuronal damage despite cessation of convulsive activity [15]. However, there is also data suggesting NCSE does not cause neuronal injury. NCSE tends to have lower-frequency discharges, which, if reproduced in animal models, cause much less neuronal damage than higher-frequency discharges. Also, animals that were previously on AEDs or had history of epilepsy are resistant to chemoconvulsant-induced damage conferring potential neuroprotection [9].

Despite conflicting evidence on molecular level, NCSE clinically has been associated with increased metabolic demand and blood flow which may cause injury to the brain. NCSE is most often secondary to an acute precipitant, which may be the underlying etiology for neuronal damage and additional injury due to NCSE may be negligible. Although human data of NCSE causing neuronal damage have been confounded by etiology, concomitant illness, and treatment, there is some evidence to suggest that there is ongoing neuronal injury secondary to NCSE [11]. For example, patients with NCSE have been found to have elevated serum levels of neuron-specific enolase, which is a marker of acute neurological injury. Patients with traumatic brain injury and subsequent NCS were noted to have delayed, prolonged increase in intracranial pressure and lactate/pyruvate ratios indicating compensatory mechanisms for increased metabolic demand [16].

The treatment of NCS and NCSE has been extrapolated from treatment of GCSE; however, the appropriateness of this practice is unclear. GCSE is a life-threatening emergency and is usually treated with high doses of sedating medications, and patients often require intubation. This aggressive approach has also been used in NCS and NCSE but adverse outcomes have been noted. Aggressive treatment of NCSE in elderly has been associated with an increased risk of death [17]. Aggressive ICU management of patients with NCSE has also been shown to prolong hospitalization without improving outcome [18]. Recent studies suggest that the use of anesthetic agents to treat NCSE may lead to higher morbidity and mortality [19, 20, 21]. In children, NCSE, but not NCS, appears to increase mortality and worsen cognitive status [22, 23].

Based on the likely differences in the pathophysiology of NCS, NCSE, and GCSE and the realization that aggressive therapy is not without potential complications, treatment of these conditions should be differentiated. Because of its recognized morbidity and mortality, GCSE should be treated aggressively. However, NCSE may not require the same degree of aggressive treatment, and NCS treatment should be even further tempered. A recent survey of neurologists noted such an approach to treating NCSE and NCS [21]. This study noted that most neurologists used nonsedating AEDs more often and were less willing to intubate for NCS as compared to NCSE.

Recognizing that the principle electrographic difference between NCS and NCSE is the amount of epileptic activity, a clinical tool, the seizure burden score (SBS), is being investigated [24]. The SBS is a composite score taking into account the amount of epileptic (ictal) time per hour weighted by the spatial extent of the seizures and EEG frequencies involved. A higher SBS score implies a more severe burden of epileptic activity. Initial reports suggest that there is a trend toward higher odds for poor outcome for patients with scores above the median compared to those with scores below the median [24].

Another way to approach treatment of comatose patients noted to have NCS and NCSE is to consider the etiology of their comatose state. If the coma is primarily due to NCSE, it is referred to as “NCSE proper,” whereas if there is another underlying etiology accounting for the coma (and possibly the subsequent seizures), it is called “comatose NCSE.” NCSE proper may be accompanied by subtle clinical symptoms suggestive of SE and mild impairment of consciousness as seen in TASE or CPSE. Comatose NCSE does not exhibit any clinical signs of SE but has characteristic epileptiform EEG patterns. The distinction between NCSE proper and comatose NCSE may be of value in determining the aggressiveness of treatment since comatose NCSE has reasons other than the SE for neuronal injury. Even aggressive treatment of the NCSE may not result in meaningful improvement and may not be indicated [20]. However, in NCSE proper, the mental status impairment and neuronal injury (if any) are due to the NCSE itself. Treating this more aggressively may result in a favorable outcome. As noted previously, a recent study provides class III evidence that therapeutic coma may be associated with poorer outcome after SE and portends higher infection rates and longer hospitalizations [25].

The treatment options discussed here have been used in NCS and NCSE. The absolute or comparative efficacy of these AEDs in NCS and NCSE is not clear as most studies are retrospective and often in patients with GCSE. Prospective studies are underway but results of these are not yet available. The discussion below pertains to the use of these AEDs in NCS and NCSE. Treatment of GCSE is likely to be different. A summary of the typical dosing, adverse effects, and limitations is presented in Table 2.