EEG forms the cornerstone for the diagnosis of NCSE and for monitoring treatment response. The vast majority of seizures recorded in critically ill patients are unrecognized at the bedside and can only be diagnosed with EEG [22]. As a general rule, any fluctuating or unexplained alteration in behavior or mental status warrants an EEG to evaluate for the presence of NCSE. EEG should be a routine component in the management of the following clinical situations:

Various criteria for which EEG patterns constitute NCSE have been proposed and modified over time [4, 10, 91]. Based on these criteria, an expert panel at the 4th London Innsbruck Colloquium on Acute Seizures in Salzburg, Austria proposed the working criteria for NCSE known as the Salzburg Consensus Criteria for NCSE (SCNC). Primary criteria for NCSE include prolonged epileptiform discharges faster than 2.5 Hz. For patterns 2.5 Hz or slower, but at least 0.5 Hz, there are secondary criteria that must be satisfied to diagnose NCSE: (1) subtle clinical ictal phenomena, (2) typical spatiotemporal evolution of rhythmic and/or epileptiform discharges, or (3) clinical and EEG response to ASD treatment. There is currently no consensus on the minimum duration of an ictal pattern to qualify as NCSE, so the distinction between recurrent NCSzs and NCSE in patients with critical illness or coma can be somewhat arbitrary. Classical coma patterns such as diffuse polymorphic delta activity, spindle coma, alpha/theta coma, low output voltage, or burst–suppression do not represent NCSE.

There are currently no randomized studies upon which to base treatment decisions for NCSE, leading to considerable controversy about whether to treat NCSE as aggressively as one treats convulsive SE [92]. In all patients with NCSE, an effort should be made to treat seizures as quickly as possible, but with minimal sedation, to avoid worsening an encephalopathy. The primary goal should be avoidance of harm that can be associated with sedation and intubation, as these risks may be higher than the risk of neuronal injury from NCSE itself. Unfortunately, with the lack of clear data to predict the risk of neuronal injury from NCSE, treatment decisions must ultimately be made on a case-by-case basis. Nevertheless, the Neurocritical Care Society [6] and the European Federation of Neurological Societies [49] have published treatment guidelines recommending that NCSE be treated similar to convulsive SE, but with additional attempts at non-coma-inducing treatment before moving to the use of anesthetic drugs. Initial treatment strategies include simultaneous assessment and management of airway, breathing, circulation (obtain IV access, administer O₂, and secure the airway as needed), abortive ASD treatment, screening for underlying causes including identification of causative drugs, and immediate treatment of life-threatening etiologies (e.g., meningitis, intracranial mass lesions, and hypertensive crisis) (Table 24.3). Importantly, C-EEG monitoring should always be utilized to guide therapy.

Unlike convulsive SE, NCSE is a more heterogeneous diagnostic entity and less amenable to standard treatment algorithms. Treatment must be individualized to the patient, tailored to the perceived urgency and morbidity of the underlying condition.

For the initial treatment of NCSE in adults, a benzodiazepine (lorazepam, diazepam, or midazolam) should be administered intravenously (or intramuscularly if peripheral access is not obtainable) in combination with a non-coma-inducing ASD for maintenance therapy, such as fosphenytoin/phenytoin, valproate, levetiracetam, or lacosamide. Suggested doses are provided in Table 24.4.

Support for the use of benzodiazepines and other ASDs for NCSE is primarily extrapolated from data in patients with GCSE. In clinical trials, the use of benzodiazepines for convulsive SE is associated with shorter seizure duration and a lower risk of cardiorespiratory complications compared with placebo [93]. There are, however, observational data to suggest that overtreatment (defined as using doses >130% of the recommended doses) might be associated with a higher need for intubation [94].

Fosphenytoin/phenytoin is the most commonly used IV ASD for NCSE [95, 96] but several studies indicate that valproate is a good alternative, with equal or better efficacy, as shown in clinical trials of patients with GCSE [96–98]. Levetiracetam is a safe alternative with few side effects and no drug interactions, but clinical trial data for its use in either NCSE or GCSE are lacking, and early evidence suggests it may be less effective than either phenytoin or valproate [99]. Lacosamide is also well tolerated and a promising alternative, but data are still limited. In 2013, a review of published case series utilizing lacosamide for the treatment of all types of SE found it effective in 56% of patients [100]. More recently, the Treatment of Recurrent Electrographic Nonconvulsive Seizures (TRENdS) study aimed to compare IV lacosamide with fosphenytoin for treatment of nonconvulsive seizures not meeting criteria for SE (defined as less than 30 min of electrographic seizure activity per 1 h of C-EEG monitoring) [101]. Patients were randomized to treatment with a loading dose of lacosamide (400 mg) or fosphenytoin (20 mg/kg), with rebolus of the ASD if breakthrough seizures occurred within 8 h of the initial bolus. Control of seizures for 24 h was considered successful treatment. Of the 62 patients who completed the study, 63% receiving lacosamide became seizure-free compared to 50% treated with fosphenytoin, indicating that lacosamide and fosphenytoin were essentially equally efficacious.

Refractory status epilepticus (RSE) refers to SE that continues despite administration of therapeutic doses of a benzodiazepine as well as a traditional ASD. At this stage, treatment of GCSE typically involves anesthetic agents, but there are few data exploring the use of anesthetic agents for refractory GCSE, let alone for NCSE. The most commonly utilized agents include propofol and midazolam due to their widespread availability and relative safety margin, although both agents commonly produce hypotension. In addition, prolonged use of propofol has been associated with the “propofol infusion syndrome” which includes metabolic acidosis, renal failure, and rhabdomyolysis, and can be fatal. Barbiturates (pentobarbital, phenobarbital, thiopental) are also a consideration at this stage, particularly when midazolam and propofol are ineffective. The advantages of barbiturates include fewer breakthrough seizures, but often at the risk of more significant hypotension, and a longer half-life can lead to prolonged ICU stays [102]. Finally, ketamine is a nontraditional anesthetic option with a novel mechanism of action (of NMDA receptor antagonism) and a low incidence of hypotension. In a retrospective case series of 60 episodes of RSE (68% of which is NCSE), ketamine appeared to contribute to control of SE in 32%, including in 12% when it was the last drug added [76].