Super refractory status epilepticus (SRSE) is defined as continuous or recurrent seizures without normalization of consciousness lasting for 24 h or more despite administration of an intravenous (IV) anesthetic (midazolam, propofol, ketamine, or barbiturate), or recurrence of status epilepticus (SE) on weaning of IV anesthetics (1). In a prospective study, 9% (12/128) of SE episodes failed second-line antiepileptic drugs (AEDs) and required coma induction. Persistence or recurrence of seizures after 24 h following treatment were not mentioned in this study; hence some of these patients may have not fulfilled the proposed criteria of SRSE (2). The term SRSE was introduced in 2011 in London-Innsbruck Colloquium; therefore, there is paucity of information on SRSE (1). SRSE is more likely following acute brain injury rather than chronic epilepsy and is more likely after central nervous system (CNS) infection, stroke, or in a healthy brain with new onset refractory status epilepticus (3). The predictors of refractory status epilepticus (RSE) are acute SE etiology, coma/stupor, and serum albumin <3.5 g/l (4). Development of RSE or SRSE was associated with stupor or coma on admission and absence of history of epilepsy. These patients had a worse outcome compared to those who responded to AEDs (5). Infections can also complicate SE and are associated with a poor outcome (6). The etiology of SRSE in the developing countries is dominated by CNS infections (7). SRSE can be convulsive or non-convulsive. The convulsive SE can result in serious complications such as myocardial injury, aspiration pneumonia, pulmonary edema, and rhabdomyolysis leading to renal failure. The diagnosis of non-convulsive SE (NCSE) requires electroencephalography (EEG), which may delay the diagnosis. SE may occur due to failure of the mechanisms responsible for seizure termination or from the initiation of mechanisms leading to prolonged seizures (t1). Persistence of SE can cause long-term sequelae (after time point t2) including neuronal injury, neuronal death, and alteration of neuronal network depending on the type and duration of seizures. For convulsive SE, t1 is 5 min and t2 is 30 min. For other types of seizures, t1 and t2 are not known (8). There is paucity of studies on SRSE, and the available information is based on retrospective review of cases. From China, 13 patients with SRSE were reported during January 2010 to August 2013 (9); from France, 78 patients were collected in an intensive care unit (ICU) survey during 2001 to 2011 having SE more than 7 days of anesthesia (10), and from South India (Hyderabad) 30 (16.9%) patients, and from Vellore 17 (7.7%) events of SRSE were reported (11, 12). In a recent population-based survey from Germany, during 2008 to 2013, 338 out of 2,585 (13%) patients with SE had SRSE (13). In the present study, we report the frequency, etiology, predictors, and prognosis of SRSE.

This retrospective analysis of patients with SRSE was conducted at a tertiary care teaching hospital in India during September 2013 to January 2016. The patient information was retrieved from a prospectively maintained SE registry. The project was approved by the Ethics Committee of Sanjay Gandhi Post Graduate Institute of Medical Sciences (N0 2017/161/IP/EXP).

Fourteen out of 107 (13.1%) patients with SE developed SRSE. The median age of the patients with SRSE was 27.5 (2–76) years; four below 18 years of age and seven were females. Metabolic etiology was present in five; acute renal failure in one, and chronic dialysis dependent in four patients. Comorbidities included hypertension in five patients, diabetes and chronic liver disease in one each. In five patients, SRSE was due to CNS infection; herpes simplex encephalitis (HSE) in two, scrub typhus in one, and nonspecific encephalitis in two. One patient had cerebral venous sinus thrombosis due to systemic lupus erythematosus. In the miscellaneous group, there were three patients; one with post-anoxic SE, one with posttraumatic arteriovenous fistula which was treated with local onyx glue injection, and another had growth retardation with hyper IgE syndrome. In the four etiologic groups of SRSE, the demographic features and duration of seizure before starting treatment were not significantly different. More than one comorbidity was present in four (28.6%) patients. The MRI findings in the encephalitis group revealed characteristic frontotemporal involvement in the patients with HSE. In the metabolic group, MRI was abnormal in two patients. In three patients, the MRI changes were attributed to SE per se.

In the present study, 13% patients with SE had SRSE and 43% of them died. The commonest etiology of SRSE was metabolic encephalopathy and CNS infections. None of the patients with SRSE had history of epilepsy. The patients in the present study have been included based on predefined inclusion criteria. In the Chinese study, there were two deaths in the hospital and four (36.6%) at 3 months. Old age and comorbidities predicted the poor outcome (6). In the French ICU survey on prolonged SRSE, out of 130 ICUs, 15 reported no patients, 19 had at least one patient fulfilling the inclusion criteria, and the remaining ICUs provided no information. The results were based on 78 patients (7). This study highlights the difficulty and limitations in obtaining information on SRSE. In the studies form South India, 16.9% and 7.8% patients had SRSE during 8 and 5 years, respectively. SRSE was significantly more common in the extremes of ages and was due to acute symptomatic etiology in 63.4% patients. The patients with encephalitis had a high chance of SRSE (8). The etiology of SRSE may have regional variation. In the study from China, 9 out of 13 patients were due to encephalitis; toxoplasma and cytomegalo virus in one each, and the etiology in the remaining patients was not confirmed. In the French study, there was diverse etiologies including epilepsy in 11 (14.1%), stroke in 10 (12.0%), metabolic encephalopathy in 5 (4.5%), encephalitis in 3 (3.0%) and neurodegenerative disorders, drug abuse, and post-surgery in 4 (5.1%) each. Categorizing SRSE on the basis of ILAE classification did not offer much advantage for general physician. Categorizing to potentially fatal or treatable disorders, however, may be advantageous.

Paucity of chronic epilepsy as a cause of SRSE has been highlighted in an earlier report (9). It is possible that the patients with epilepsy are under treatment and surveillance, which may account for their seizures to be treated better compared to those with new onset seizure. This was noted in our study as well as studies from China and South India (6, 8). The median duration of SE in our study was longer in encephalitis (240 h) compared to miscellaneous (168 h) and metabolic encephalopathy (48 h) groups. The patient with cerebral venous sinus thrombosis had SE for 432 h. The duration of MV in our study was longer in miscellaneous compared to metabolic and encephalitis groups. In the study from France, the median ICU stay after administration of general anesthesia was 17 days (7). In the study from South India, the mean duration of ICU admission and duration of MV was longest in children (8). In spite of such long duration of SE with higher frequency of MV, the mortality in our study was comparable with other reported series (43%) (Table 5).

In view of poor outcome in the patients with SRSE, ethical concerns have been raised on the cost effectiveness of ICU care in these patients (7). However, the recent reports have suggested that acceptable recovery is possible following aggressive treatment. In our study, the encephalitis patient especially those with a treatable etiology survived. On the other hand, SRSE patients with irreversible etiology such as dialysis-dependent uremic encephalopathy had poor outcome. It is important to identify the patients who will respond to aggressive therapy. The prognostic predictors in the retrospective studies included male gender, younger age, burst suppression within 7 days, treatable etiology, and lack of need of vasopressors. The ICU-related complications included hypotension, pneumonia, sepsis, and cardiac arrhythmia, and may contribute to the poor outcome. Hypotension was seen in 92.9% of our patients; however, it did not adversely affect the outcome.

Our study is limited by small sample size and high proportion of metabolic and infectious etiologies. We did not do continuous EEG monitoring in all the patients, which may have underestimated NCSE in our study. We have used burst suppression in one patient only. However, concern has been raised on the safety and efficacy of IV anesthetic drugs in SE. In a retrospective review of 171 patients, 37% were treated with IV anesthetics, which resulted in more infections (43 vs 11%) and 2.9-fold relative risk of death (20). In another study on 126 patients with SE, administration of IV anesthetic was associated with poor outcome (21). These studies raise the question whether IV anesthetic drug do more harm than good, and should we change our practice? (22). There is a need of a prospective study on SRSE.

This study was approved by Institutional Ethics Committee, SGPGIMS, Lucknow, India.

Substantial contributions to the conception or design of the work; the acquisition, analysis, or interpretation of data for the work; and drafting the work or revising it critically for important intellectual content: UM, JK, DD. Final approval of the version to be published: UM.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.