Nonconvulsive status epilepticus following cardiac arrest—are we missing the beginning?

Cardiac arrest is the most common cause of death, resulting in up to 60,000 annual deaths in out-of-hospital cardiac arrests in Switzerland [20]. The degree of hypoxic–ischemic brain injury and dysfunction varies and can result in persistent coma or death [6, 7, 11].

For patients with persistent coma following return of spontaneous circulation (ROSC), the recommendations for initial care include artificial ventilation, sedation, and neuroprotective measures including continuous monitoring of core temperature to detect and suppress fever (defined as a temperature > 37.7 °C) for at least 72 h [4, 12, 13]. In up to one third of these patients, seizures and status epilepticus (SE) may emerge with or without clinically overt signs [1, 16, 17, 19]. While recovery from post-cardiac arrest SE has rarely been described with good outcome [5, 8, 17, 19], it has been identified as an independent predictor of poor outcome in larger cohorts [16]. In contrast to convulsive SE, nonconvulsive epileptic complications [17, 18] can be frequently missed especially without continuous EEG (c-EEG) and little is known about the frequency and effects of nonconvulsive SE (NCSE) in this context. Although in the early treatment phase of resuscitated patients with persistent coma, detection of NCSE before the initiation of iso- or hypothermia is not part of most treatment protocols, it seems plausible that unrecognized and untreated ongoing SE may cause subsequent neurologic injury that has the potential to cause further cerebral damage that may be relevant.

We aimed to compile the current evidence from the literature on the emergence of NCSE in adults with persistent coma in particular within the first hours post-ROSC and before inducing target temperature management.

We screened PubMed using the predefined search terms “status epilepticus,” “nonconvulsive status epilepticus,” “continuous electroencephalogram,” “continuous EEG” in combination with the term “cardiac arrest,” “postanoxic,” “ischemic-anoxic.” Papers in a language other than English, papers not describing SE semiology, pediatric studies, and animal models were excluded. Prospective or retrospective studies and case series reporting the frequency of NCSE were included. Data regarding study design, number of patients, time of EEG performance, number of patients with SE, clinical signs of SE, proportion of patients with early NCSE, reported outcomes of patients with NCSE, and targeted temperature management were assessed.

Detailed information of three identified prospective and retrospective cohort studies reporting the frequency of NCSE in the first hours during ongoing sedation and target temperature management is compiled in Table 1. Studies targeting explicitly the presence of NCSE in resuscitated adult patients with persistent coma after ROSC and prior to the initiation of targeted temperature management could not be identified by our review of the literature.

Overall, the three studies examined a total of 213 patients with SE in 18–38% and NCSE in 5–12% of patients.

The first and only study with a clear focus on NCSE in adult post-cardiac arrest patients was performed retrospectively and published by Rittenberger et al. in 2012 [15]. In that study, NCSE was detected in 12% of patients (12/101) using a 22-channel c‑EEG recording at a median of 9 h (range: 6–12 h) from cardiac arrest. Patients were already under targeted temperature management with a mean temperature of 33.9 °C and sedated for 48 h. Nonconvulsive status epilepticus was detected at the onset of the recording in three of the 12 (25%) patients and had a median duration of 5 h. The outcome was poor with only one patient (1%) surviving with a Cerebral Performance Category score (CPC) of 1.

Nonconvulsive status epilepticus was reported in two other studies on predictors of awakening from post-ischemic–hypoxic SE [17] and on frequency and timing of epileptiform activity on c‑EEG in comatose patients [10]. The first study (consisting of one large prospective and one retrospective part) comprising a total of 181 patients (with 74 patients investigated prospectively) revealed SE in 38% of the prospectively examined patients (28/74). The EEG was carried out using 14 or 21 channels. Data regarding EEG timing and duration of monitoring were not provided. The reported semiology was myoclonus in 31%, prolonged tonic–clonic seizures in 1%, and no motor symptoms in 5% [17]. While all but one of the patients having SE with motor symptoms died during their hospital stay, one patient remained in a minimally conscious state before dying in hospital and one significantly improved (i.e., CPC score of 2) [17].

Another retrospective study examined 38 comatose patients treated with therapeutic hypothermia and monitored with a 16–18 channel c‑EEG that was applied within a median of 15 h from arrest (range: 7.5–21.3 h) [10]. This monitoring detected SE in 23% of patients (7/38) with 13% (5/38) having SE with motors symptoms and 5% (2/38) being in NCSE. All these patients with NCSE died in hospital.

Our review of the literature identified only three cohort studies of different designs, limited quality, and with very limited information regarding the emergence of early NCSE following cardiac arrest and associated outcomes.

Any studies on the emergence of NCSE in resuscitated adult patients remaining in coma within the first hours of ROSC and prior to the initiation of targeted temperature management were identified.

The fact that only very few data were identified linking the emergence of NCSE in the early stage of post-ROSC management to a specific presumed outcome seems unjustified and critical. In addition, the data compiled in this review do not allow any conclusions to be drawn on the effects of early detection and aggressive treatment of NCSE on outcome, nor on whether an early NCSE can be only a transient condition with a different clinical impact than SE occurring at a later stage.

With the exception of the retrospective study by Rittenberger and colleagues in 2012 that used a 22-channel continuous EEG recording early after cardiac arrest and focused on the frequency of NCSE [15], the studies did not perform EEGs in the early stage and were not primarily designed to investigate the emergence of NCSE in this context. Furthermore, the study of Rittenberg et al. [15] had limitations regarding the diagnosis of NCSE. Although the authors stated that SE was considered nonconvulsive if no motor symptoms were seen on the recorded video or at clinical examination, they had to admit that only a minority of individuals had such video recordings during EEG. Moreover, one NCSE patient was described as having subtle motor symptoms, not strictly fulfilling the criteria for NCSE.

Overall, the EEGs were rarely performed very early across studies (the earliest was at a median of 7 h), in very different time windows among the studies, and sometimes the time at which the EEG was started was not reported. Moreover, the fact that EEG, even if performed early, was always performed on patients already under hypothermia and ongoing sedation is a major limitation, as both hypothermia and sedatives may have antiseizure effects. Therapeutic hypothermia per se has some antiseizure effects [14] and often requires the administration of anesthetics, such as midazolam or propofol, which may further suppress seizures. Moreover, hypothermia can increase the blood concentration of propofol [9] and its anticonvulsant effects. To what degree established transient hypothermia and sedation in the context of targeted temperature management terminate seizures despite not have been titrated to a stable burst-suppression for 48 h, as recommended in cases of treatment-refractory SE by the International League Against Epilepsy (ILAE; [3]) remains unanswered. Hence, we cannot exclude that patients with missed early NCSE would benefit from supplemental antiseizure medication in addition to antiseizure effects from hypothermia and anesthetics regarding seizure termination and outcome. The concept of a “three-dimensional” biological continuum clearly shows the complexity in the interrelation between structural brain damage, epileptic activity, and the degree of coma, and how an excessive epileptic activity can play an important role in the worsening of the brain damage itself [2].

Studies on nonconvulsive status epilepticus (NCSE) immediately after return of spontaneous circulation (ROSC) in adults and prior to the initiation of targeted temperature management are lacking and the few studies on NCSE following ROSC are unsystematic and of low quality. These findings likely show that NCSE prior to targeted temperature management and sedation is often missed. Further studies are urgently needed to uncover the true frequency of NCSE and its effect on outcome in this context and to clarify whether rigorous initiation of aggressive antiseizure treatment may be key for better recovery. In the meantime, we suggest withdrawing sedation once patients are admitted to the intensive care unit and with persistent coma to start EEG monitoring to detect or exclude NCSE. This procedure, however, should be postponed in favor of the initially necessary and more vital support interventions. The complete neurological prognostication assessment should then strictly follow current international guidelines.