There is abundant evidence from clinical studies and animal models as to the deleterious effect of on-going seizure activity in the context of CSE [3, 5, 18, 19], necessitating prompt intervention and treatment [1]. In contrast, there is much debate concerning the evidence for neuronal damage in NCSE [15, 16]. Moreover, the decision to treat is based on the risk–benefit ratio, whereby the efficacy of an intervention has been firmly established and the benefits outweigh any potential risks or the risk of nonintervention [9]. Ultimately, the decision whether treatment is initiated or not is predicated on whether a reliably effective treatment exists and whether failure to treat results in on-going deleterious cerebral effects and consequently increased morbidity and mortality. We will examine the evidence for efficacy and necessity of treatment for the four main individual subtypes of NCSE, namely typical absence status epilepticus, simple partial status epilepticus, CPSE, and NCSE in coma. Lastly, we will address the issue of use of anesthesia in NCSE.

Typical absence status epilepticus (ASE) usually occurs in children or adolescents with an idiopathic generalized epilepsy (IGE) or those with a history of an IGE in childhood; it can, however, sometimes occur de novo or as a late presentation of IGE in the elderly [20]. It manifests as varying degrees of alteration in mental status or behavior. This is associated with the classical EEG pattern of large symmetrical bilateral synchronous 2.5–3 Hz spike and wave discharges, although this pattern may become discontinuous the longer it persists. The natural history of typical ASE is a duration of several hours, with spontaneous cessation, although it can persist several days, ending in a generalized convulsive seizure in up to a third of cases. Typical absence status is often precipitated by anti-seizure drug (ASD) withdrawal, or ASD non-adherence—with re-instatement of therapy a significant component of management. Typical ASE can also be precipitated by the use of inappropriate ASDs such as carbamazepine or phenytoin [21].

There are several genetic models of absence epilepsy and seizures in rodents, such as the homozygous tottering (Cacana1atg) model and the homozygous lethargic (Cacnab41h) model, each manifesting with spontaneous frequent absence seizures resembling continuous ASE, in addition to the pentylenetetrazole (PTZ) model of pharmacologically induced absence seizures [22–24]. Importantly, none of these models exhibits neuropathological sequelae [24]. Similarly, there appears to be no clinical evidence of increased morbidity or mortality associated with ASE in children [25–27], nor indeed in de novo ASE in later life [28], which can often occur as a consequence of benzodiazepine withdrawal [29]. Indeed, continuous generalized 2.5–3.5 Hz spike wave discharges were identified on EEG in one 8-year-old girl followed up over 9 years without neurocognitive sequelae [30].

Acute treatment of ASE is, therefore, mostly aimed at preventing injury due to the altered conscious state or the occurrence of a convulsion. The consensus is that ASE responds rapidly to treatment with oral or IV benzodiazepines [6, 7, 9, 27] particularly clonazepam [27] (Table 22.2 [10, 31]), or IV valproate or oral acetazolamide when benzodiazepines are contraindicated. Moreover, successful management requires identification and correction of and avoidance of potential precipitating factors, if possible, such as acute systemic infections, psychotropic drugs, use of inappropriate ASDs, or alcohol withdrawal [26, 27, 29, 32]. Interestingly, episodes of ASE may be triggered by use of inappropriate ASDs such as carbamazepine and phenytoin which may render standard treatment [benzodiazepines (BZDs) or valproic acid (VPA)] ineffective due to a paradoxical effect of the inappropriate ASD, underscoring the risk of iatrogenic exacerbation of ASE [21, 33]. For people with IGE prone to recurrent episodes of typical ASE, long-term therapy appears to be efficacious in significantly reducing or preventing further attacks [34]. The elderly with NCSE de novo rarely require longer term treatment.

Atypical ASE, which typically occurs in the context of the epileptic encephalopathies such as Lennox–Gastaut syndrome, can be quite refractory to non-anesthetic drugs, which nevertheless remain the preferred method of treatment. Moreover, treatment with BZDs needs to be employed cautiously, as this may precipitate tonic SE in some of these patients [35].

Compared to typical ASE, impairment of consciousness is reputedly more severe in atypical ASE, but the EEG may not differentiate the two. Treatment with second-line agents such as VPA or phenobarbital may be necessary at times (Table 22.2). Anesthesia should be avoided as treatment for patients with ASE (see below).

Simple partial status epilepticus (SPSE) or aura continua is considered a rare form of SE, but because it lacks motor features or impairment of consciousness, it may go unrecognized. SPSE arises as a result of seizure activity involving a focal area of cortex, with diverse clinical presentations including aphasia, visual or auditory hallucinations, altered perception, transient visual loss, and episodes of fear. Similarly, SPSE may take the form of focal motor seizures, aversive eye movements, transient paralysis contralateral to the epileptic focus, or focal myoclonic jerks with or without impairment of consciousness (epilepsia partialis continua [EPC]) lasting days to months—which may be highly refractory to treatment [36]. In the largest reported series, among 47 people with new onset SPSE identified retrospectively over a 5-year period in the Netherlands, all but one (with aphasia) had somatomotor SPSE; 20 had EPC. Just over half (27) had a previous diagnosis of epilepsy, although the SPSE was caused by a new neurologic diagnosis in six. Of those without a prior diagnosis of epilepsy, cerebrovascular disease was the most common etiology of SPSE, in 14 patients. Overall, four patients died (all due to acute cerebrovascular disease) and ten had associated morbidity, only one case of which was felt to be attributable to the SPSE. In general, SPSE appears to be associated with low mortality and morbidity, almost always caused primarily by the underlying etiology. Treatment, first with oral or IV BZDs (e.g. clonazepam) followed, if necessary, by IV phenytoin, is usually effective (Table 2). Anesthesia should, if possible, be avoided in patients with SPSE (see below).

Complex partial status epilepticus (CPSE) is in contrast probably the most frequent subtype of NCSE, reportedly accounting for up to 40% of cases of SE overall [12, 37]. The presentation of CPSE can be very heterogeneous and primarily determined by the origin of the epileptic focus. Fluctuating consciousness, with or without amnesia is the typical presentation, but altered consciousness may not occur in CPSE with a unilateral frontal lobe focus [38]. It can be difficult to distinguish CPSE from ASE, especially in the elderly with de novo ASE. Moreover, while CPSE needs to be differentiated from other forms of NCSE, it also needs to be distinguished from other causes of encephalopathy (e.g., hepatic encephalopathy), prolonged post-ictal states, or primarily psychiatric behavioral changes. This is often not straightforward. EEG is not always helpful, as the changes on scalp EEG in CPSE can be relatively nonspecific, so the diagnosis of CPSE remains primarily an electro-clinical one. Nevertheless, strict electrographic criteria for the diagnosis of CPSE need to be employed [39, 40].

Similarly, response of the EEG to treatment, in particular with BZDs, cannot be used as a reliable diagnostic indicator of CPSE because some EEG abnormalities, such as the triphasic waves of hepatic encephalopathy, may also improve with BZDs [41]. Conversely lack of immediate clinical improvement should not exclude the diagnosis, as often the confusion of CPSE may be replaced by the confusion of the post-ictal state without any obvious change in clinical condition.

CPSE is considered the result of focal seizures spreading to involve bilateral cerebral structures, presenting with variable degrees of impairment of consciousness, with partial or complete amnesia, typically gradual in onset and often recurrent [42]. While originally referred to as “psychomotor status” arising from the temporal lobes, a depth electrode study of 87 people with complex partial seizures showed that of the eight who developed CPSE, all had extra-temporal CPSE, with frontal lobe origin definitely established in four [38].

In a prospective study of 10 patients identified with CPSE of frontal lobe origin, there was a significant delay in recognition and diagnosis of CPSE (mean time to diagnosis, 48 h). In the first type of frontal CPSE, seven patients presented with mood disturbances or affective disinhibition with subtle impairment of cognitive function, but without overt confusion [43]. This was associated with a unilateral frontal epileptiform ictal pattern, with normal background activity. In the second type of frontal CPSE, three patients presented with impaired consciousness with an EEG pattern of bilateral asymmetric frontal discharges on an abnormal background [43].

CPSE is clinically subdivided into (1) CPSE in patients with epilepsy and (2) CPSE de novo. This is in part based on the recognition that the prognosis of CPSE in those with a prior diagnosis of epilepsy is far better than that when CPSE occurs de novo in the context of an acute medical illness. Indeed, experimental evidence indicates that epilepsy and previous exposure to ASDs may be neuroprotective in CPSE [44].

Management of CPSE is controversial, with many authors advocating aggressive treatment. This is, in part, because animal models of CPSE can demonstrate severe neuronal damage [45]. Nevertheless, the mechanisms of seizure generation and propagation in animal models may not accurately reflect human CPSE [6]. Moreover, CPSE in humans is typically characterized by lower frequency discharges than those seen in animal models [46]. When animal models are reproduced using these lower frequency discharges, significantly less damage is produced than with the higher frequency discharges [47].

The clinical data from case series suggest a more favorable prognosis, particularly in people with CPSE in the context of previous epilepsy [42, 44]. In one series, 20 patients with CPSE were identified, of whom 17 had recurrent episodes, many of which were very prolonged, lasting a few days to a few months, with one case lasting 18 months [42]. In the 17 patients with recurrent episodes, nine experienced monthly episodes, with a range of one to six episodes per month. Despite the relative frequency and longevity of the episodes, none reported or demonstrated any neurocognitive sequelae. Five of the 17 patients underwent neuro-psychometric evaluation on more than one occasion over a period of two or more years, with no evidence of cognitive decline on formal testing.

This and other case series of CPSE [38, 42–44] suggest a relatively favorable prognosis in terms of neurocognitive sequelae. CPSE in those with prior epilepsy is typically responsive to oral BZDs such as clobazam or intravenous BZDs such as lorazepam or diazepam, although recurrence is not uncommon [42]. There may, however, be a delayed clinical response despite resolution of EEG features as the patient enters a post-ictal state. Consequently, EEG monitoring, when available, should be used to monitor the response to treatment. In the case of failure of response to oral or IV BZDs, IV ASDs can be considered [48, 49] (see Table 22.2). In the scenario where CPSE with prior epilepsy remains refractory to non-anesthetic agents, treatment escalation to IV anesthetics should only be considered with caution and on an individual basis. The paucity of clinical data for long-term neurocognitive consequences underlies the reticence of many neurologists to advocate escalation to IV anesthetic agents, given their association with morbidity and mortality [50].

De novo CPSE, occurring in the context of acute or progressive neurological conditions, is in contrast, often refractory to non-anesthetic agents. Prognosis, as with generalized CSE [51], is primarily determined by the underlying etiology, and successful management of the CPSE is contingent on the rapid recognition of the precipitating cause if possible. This is particularly true in the case of autoimmune encephalitis, such as anti-NMDA receptor encephalitis, which can result in CPSE resistant to standard ASDs and IV anesthetic agents, responding only after initiation of immunosuppressive therapy (IV steroids, IVIG, or plasma exchange) [52]. Because of the potential response of these conditions to immunosuppression, autoimmune encephalitis should be considered in all patients with SE and encephalitis of unknown origin, with early use of immunosuppression when there is a high clinical suspicion. In all these cases, suitable examination and investigation should be performed to look for occult tumors (in particular ovarian, breast, and testes).

CPSE may be underdiagnosed in the ill elderly, and may contribute to confusion and the clinical state. In this instance, care needs to be taken with the use of intravenous BZDs, which can increase mortality. Generally, intravenous valproate or levetiracetam are preferred, as they have fewer adverse hemodynamic and respiratory effects. Anesthesia should, if possible, be avoided in patients with non-comatose CPSE (see below).

Nonconvulsive status epilepticus in coma is considered the most severe subtype, with significant morbidity and mortality, and often poses the greatest diagnostic dilemma. It is also the area where there is the most controversy regarding treatment, with a current state of equipoise between aggressive intervention and nonintervention. NCSE in coma can be subdivided into three groups: (1) “subtle status epilepticus” as originally conceived by Treiman (1993), referring to the end stage or burnout of overt generalized CSE, whereby NCSE evolves from generalized convulsive status epilepticus (GCSE), typically in the context of partially or nontreated GCSE [53]; (2) electrographic SE with subtle clinical signs of seizure activity at presentation; and (3) electrographic SE found incidentally on EEG in the critically ill in coma.

Despite the significant morbidity and mortality associated with NCSE in coma there are few data to guide management or indeed, to inform us if intervention even impacts overall prognosis. “Subtle” NCSE is considered the most malignant form of NCSE, the management of which was studied in the Veteran’s Affairs Status Epilepticus Study [54], where 570 patients with SE (375 with overt SE, and 175 with subtle SE) were randomized to one of four treatment arms: administration of lorazepam (0.1 mg/kg); diazepam (0.15 mg/kg) followed by phenytoin (18 mg/kg); phenobarbital (18 mg/kg); or phenytoin (18 mg/kg) alone. Successful cessation of SE was significantly better for overt SE compared to subtle SE in all four treatment arms: lorazepam (67% for overt SE, compared to 26.1% for subtle SE); phenobarbital (63% compared to 24.4%); diazepam and phenytoin (59.6% compared to 23.4%) and phenytoin (51% compared to 19.5%). While lorazepam was found superior to phenytoin in the treatment of overt SE (p = 0.002), no difference was found between any two treatment arms for the management of subtle SE (p = 0.18) [54].