The Refractory Status Epilepticus (RSE) is out of the Barn… the Current State of Refractory Status Epilepticus Management and Outcomes

Abstract

Features Affecting Treatment Decisions and Outcome in Refractory Status Epilepticus

Damien C, Torcida Sedano N, Depondt C, Legros B, Gaspard N. Epilepsia. 2025;66(8):2779-2789.

Objective: Refractory status epilepticus (RSE) is associated with worse outcomes than responsive established status epilepticus (SE). Guidelines recommend that refractory convulsive SE should be treated with continuous intravenous anesthetic drugs (CIVADs). Many cases of nonconvulsive SE are not treated with CIVADs, and the use of anesthesia might be associated with increased mortality. The factors leading to the decision to use anesthesia and how these might affect outcome are still largely unknown. Our goal was to identify features of refractory SE associated with treatment choices and outcome. Methods: A single-center, retrospective study was conducted of all consecutive patients with RSE admitted to a tertiary center between January 2015 and December 2020. We collected demographic and clinical variables at SE onset and at time of third-line treatment, including ictal burden during the hour preceding the administration of the third-line treatment. The primary outcome measure was the decision to use CIVADs as third-line treatment. Secondary outcome measures were in-hospital mortality and functional outcome at discharge. Results: One hundred sixty-one RSE episodes were included. Of these, 29 (18%) received CIVADs as third-line treatment and 61 (38%) died. The type of third-line treatment was not associated with mortality. CIVADs were more likely to be used with higher ictal burden, fewer comorbidities, a lower Glasgow Coma Scale (GCS) score at time of third-line administration, and in the absence of history of epilepsy (odds ratio [OR] = 1.03, 0.76, .66, and .25, respectively). Multivariable analyses also identified comorbidities, an acute etiology, and lower GCS score at time of third-line administration as risk factors of mortality (OR = 1.43, .09, .28, and .80, respectively). Ictal burden was not associated with outcome. Significance: Ictal burden, semiology, and consciousness at time of third-line treatment are associated with the decision to use CIVADs in SE. Semiology and consciousness at time of third-line treatment are also associated with mortality.

Commentary

Refractory status epilepticus (RSE), defined as status epilepticus that continues despite adequate doses of one first-line benzodiazepine and one appropriate second-line antiseizure medication, has a high risk of morbidity and mortality. ¹ Approaches vary between treating healthcare providers with regard to electroencephalogram (EEG) interpretation, threshold for, and comfort with starting sedating anesthetic medications to treat RSE. ²

Damien et al ³ examined a retrospective review of a prospective database of 161 adults treated for RSE at Erasme Hospital from 2015 through 2020 to assess management strategies and clinical outcomes. They applied EEG and clinical criteria to quantify ictal burden, used the Charlson comorbidity index to measure comorbidity burden, and the Glasgow Coma Scale score to assess change in neurologic function over time. Status Epilepticus Severity Score (STESS), which includes a composite of level of consciousness, seizure type, age, and seizure history at the onset of status epilepticus, was also used to predict functional outcomes.

The authors found that intravenous anesthetic agents were used third-line in only 29 (18%) of patients, and the mortality rate was 38% overall. Patients with higher comorbidity burden, worse neurologic function at the time of RSE, acute onset, and no prior history of epilepsy were more likely to be treated with anesthetics as a third-line therapy. Nonconvulsive (NC) status epilepticus at onset, higher comorbidity burden, and absence of consciousness (lower modified Rankin scale score) at initiation of an intravenous anesthetic were all associated with poor outcome. Higher STESS was also associated with poor outcome, which helps validate the use of this tool for outcome prediction. Surprisingly, ictal burden at anesthesia initiation and progression to super-RSE (RSE that persists for 24 h or more despite aggressive management or returns after attempting to wean anesthetic agents) ⁴ were not associated with worse outcome. In addition, when reviewing EEG findings prior to anesthesia initiation, 30% of patients treated did not meet the applied criteria for RSE in the 1 h prior to initiation.

The authors illustrated important study limitations, most notably its retrospective nature. This is particularly true when considering the latter finding that 30% of EEGs did not meet the criteria for RSE before anesthesia initiation. In a real-world scenario, the decision to start anesthetic agents may take place several hours before the initial dose is started, and therefore, evaluating based on only the 1 h immediately preceding initiation may not reflect the EEG findings at the time this decision was reached. Expert EEG interpretation has also been found to be highly subjective, with poor inter- and, at times, intra-rater reliability as well. ⁵ These observations highlight the dynamic nature of the disease process and difficulty in determining the best time to escalate treatment. The finding that NC RSE was associated with worse outcome supports the need for rapid initiation of long-term, continuously monitored video EEG in order to immediately diagnose and treat this condition, and if unavailable, to immediately transfer a patient to a facility with these capabilities. Intermittent EEG monitoring could delay NC RSE discovery, appropriate treatment escalation, and result in worse outcomes and death.

A subset of patients in the study were conscious when anesthesia was administered, which is a challenging circumstance that healthcare providers, patients, and families can have a difficult time trying to navigate. For example, an asymptomatic patient with EEG findings meeting criteria for electrographic status epilepticus or a patient with epilepsia partialis continua (continuous electroclinical motor seizures with maintained consciousness) ⁶ may be resistant to consider aggressive therapy. In these cases, the decision regarding whether or not individuals are started on anesthesia can be highly dependent upon the etiology, patient and family preferences, as well as the comfort and experience of the treating team with managing this condition. The optimal approach is unknown, and the study findings in this population may not be easily generalizable to other medical centers.

There are several notable limitations with regard to the study population described. The database from which patients were selected included nearly one year during the COVID-19 pandemic, a time which resulted in healthcare avoidance among patients, a reduction in healthcare resources, increased psychosocial stress, and higher non-COVID death rates overall. ⁷ It would be interesting to investigate whether this impacted disease severity and comorbidities in patients with RSE, clinical practice in RSE management, and overall outcomes compared to the 5-year interval prior. Finally, individuals with post-anoxic RSE were excluded from the investigation. Studies have previously shown that there is a higher rate of morbidity and mortality in this patient population. However, outcomes may improve with aggressive management ⁸ and therefore, inclusion of these patients in future studies is warranted.

Overall, the findings of this study support the need for robust long-term EEG monitoring resources to diagnose and treat RSE. The study provides an excellent road map for investigators designing clinical trials to treat RSE when making power calculations based on anticipated outcomes and identifying potential outcome measures to utilize. Innovative and evidence-based treatment strategies are needed to ultimately lead to reductions in morbidity and mortality based on the type and etiology of RSE, as well as patient comorbidities, tailored to the individual and circumstance.