Cryptogenic new-onset refractory status epilepticus responded to anti-interleukin-6 treatment

OBJECTIVE
The aim of this case report was to describe a potential anti-interleukin (IL)-6 treatment for cryptogenic new-onset refractory status epilepticus (C-NORSE).


BACKGROUND
Although an underlying immune-mediated pathogenesis is considered present in some C-NORSE cases, many cases do not respond to classical immunotherapies.


CASE REPORT
We describe the case of a 46-year-old woman with C-NORSE who achieved cessation of long-lasting status epilepticus following administration of tocilizumab, an IL-6 receptor-blocking antibody, although the final outcome was poor.


CONCLUSIONS
Anti-IL-6 treatment may prove effective in stopping status epilepticus in some C-NORSE cases.


Introduction
New-onset refractory status epilepticus (NORSE) describes very longlasting status epilepticus (SE) without a readily identifiable cause in otherwise healthy individuals. NORSE was first described by Smith in 2007(Wilder-Smith et al., 2005. Despite extensive research efforts, around half of adult NORSE cases remain cryptogenic (Sculier and Gaspard, 2019). Although several studies have reported cryptogenic NORSE (C-NORSE) cases in which classical immunotherapies, such as high-dose corticosteroids, intravenous immunoglobulins (IVIg), or plasma exchange (PE), proved effective (Gall et al., 2013;Kodama et al., 2018), many C-NORSE cases do not respond to such therapies, and display poor outcomes (Iizuka et al., 2017).
Here we describe a case in which cessation of SE was successfully achieved using tocilizumab, an interleukin (IL)-6 receptor-blocking antibody, for C-NORSE refractory to high-dose corticosteroids, IVIg, PE, and intravenous cyclophosphamide (IVCY).

Case report
The patient was a 46-year-old woman with no significant past medical history. She was admitted to a previous hospital with new-onset generalized tonic-clonic seizures two weeks after developing headache and one week after developing fever. The patient was intubated and placed on mechanical ventilation immediately after admission. Although several oral antiepileptic drugs (AEDs), propofol, and midazolam were used, she experienced repeated convulsions of the face and right limbs for 3 days, and was transferred to our hospital.
On admission, her level of consciousness was low (Glasgow Coma Scale score 3; E1VTM1) on high-dose propofol and midazolam, and she still displayed frequent facial twitches on either side, followed by generalized convulsions. Blood tests did not reveal any metabolic disorders, and results for serum autoantibodies were negative except for anti-thyroglobulin antibody. Cerebrospinal fluid (CSF) examinations revealed elevated cell counts (40/μL). However, other general tests were normal, and results from bacterial and fungal cultures, DNA polymerase chain reaction assays for herpes simplex virus and varicella-zoster virus were negative. Autoantibodies against neuronal antigens including N-methyl-D-aspartate receptor (NMDAR), leucine-rich glioma-inactivated 1, contactin-associated protein-like 2, α-amino-3-hydroxy-5-methyl-4isoxazolepropionic acid receptor, γ-aminobutyric acid A receptor, γ-aminobutyric acid B receptor, glutamic acid decarboxylase, metabotropic glutamate receptor 1, metabotropic glutamate receptor 5, dipeptidyl-peptidase-like protein 6, neurexin-3, and immunoglobulinlike cell adhesion molecule 5 were all negative. Indirect immunofluorescence using serum from the patient and a rat brain section revealed negative signals. Fluid-attenuated inversion recovery magnetic resonance imaging (MRI) of the brain revealed high-intensity areas in bilateral medial temporal lobes and insulas (Fig. 1). Magnetic resonance angiography showed no abnormalities (including mycotic aneurysms), and transthoracic echocardiography showed no abnormalities (including vegetations or valvular diseases). Computed tomography of the trunk and MRI of the abdomen revealed no tumors. Since the patient did not have any evident cause of SE, including cerebrovascular disease, central nervous system infection, brain tumor, metabolic encephalopathy, toxic encephalopathy, or already-known autoimmune encephalitis, C-NORSE was diagnosed.
After admission, convulsions were controlled using high-dose thiamylal (maximum dose: 5.6 mg/kg/h), and electroencephalography (EEG) of the patient revealed a burst suppression pattern ( Fig. 2A). However, whenever we reduced doses of thiamylal, similar convulsions immediately recurred, and EEG revealed frequent sharp waves mainly observed bilaterally in the parieto-occipital area, and an ictal EEG pattern of focal seizure several times during the recording (Fig. 2B). Other intravenous anesthetics including propofol (maximum dose: 3.5 mg/kg/h), midazolam (maximum dose: 0.3 mg/kg/h), and ketamine (maximum dose: 2.4 mg/kg/h) were administered, but all proved ineffective, with no resolution of convulsions or epileptic discharges on EEG. We thus had to use high-dose thiamylal for control of SE over a long period of time. Many oral AEDs, including phenytoin (maximum dose: 600 mg/day), phenobarbital (maximum dose: 360 mg/day), carbamazepine (maximum dose: 1200 mg/day), levetiracetam (maximum dose: 1500 mg/day), lacosamide (maximum dose: 50 mg/day), sodium bromide (maximum dose: 1500 mg/day), perampanel (maximum dose: 12 mg/day), and topiramate (maximum dose: 600 mg/day), were trialed for control of SE, but all again proved totally ineffective, or caused several adverse events (Fig. 3). A ketogenic diet was trialed for almost one month, but convulsions and epileptic discharges on EEG were not resolved at all, and the diet had to be discontinued due to severe diarrhea.
Next, we planned to use tocilizumab, an IL-6 receptor-blocking antibody, considering that IL-6 in CSF appeared significantly high (118 pg/mL; no cut-off value has been confirmed, but a previous study suggested a cut-off of 10 pg/mL for non-herpetic acute limbic encephalitis (Ichiyama et al., 2008)) on day 137 of admission. After gaining approval from the Osaka University Clinical Research Review Committee, we administered tocilizumab on days 141 and 148 (4 mg/kg each). The frequency of convulsions gradually decreased, and we were mostly able to control convulsions using midazolam without thiamylal. We administered tocilizumab a third time (8 mg/kg) on day 187, finally achieving total control of convulsions in combination with four AEDs (perampanel at 12 mg/day, topiramate at 600 mg/day, clobazam at 15 mg/day, and lamotrigine at 300 mg/day) and no intravenous anesthetics. EEG revealed diffuse, low-amplitude slow waves, and did not show any epileptic discharges after administrations of tocilizumab on day 246 (Fig. 2C). The patient was able to be withdrawn from mechanical ventilation on day 220. IL-6 in CSF decreased markedly and was 6.4 pg/ mL after the last administration of tocilizumab on day 206. Since convulsions stabilized, she was transferred to another long-term care hospital in a state of unresponsive wakefulness syndrome (modified Rankin Scale (mRS) score 5) on day 280 of admission ( Fig. 3).

Discussion
We encountered a 46-year-old woman with C-NORSE who achieved cessation of long-lasting SE following 3 administrations of tocilizumab, although the final outcome was poor.
The etiology of NORSE is either unknown or an unusual cause only identified after extensive work-up, and can be divided in 4 categories: inflammatory and autoimmune (the most frequently identified cause; 40% of NORSE cases), uncommon infectious encephalitis, genetic disorders, and toxic disorders (Sculier and Gaspard, 2019). Around half of adult NORSE cases remain cryptogenic, despite extensive work-up (Sculier and Gaspard, 2019). C-NORSE may represent a heterogeneous group of disorders, but the clinical features are often similar to those of autoimmune cases, so some C-NORSE cases may correspond to autoimmune encephalitis associated with unidentified autoantibodies or not Fig. 1. Findings from MRI of the brain. Fluid-attenuated inversion recovery images reveal high-intensity areas in bilateral medial temporal lobes and insulas on day 37 of admission (arrowheads).

Fig. 2. EEG findings.
(A) When convulsions are controlled using high-dose thiamylal on day 9 of admission, EEG reveals a burst suppression pattern. (B) On reducing the dose of thiamylal on day 89 of admission, the patient displays frequent facial twitches on the right or left side, followed by generalized convulsions. EEG the same day reveals frequent sharp waves, mainly observed in bilateral parieto-occipital areas. An ictal EEG pattern is also observed: rhythmic β waves start at the right central area (arrowhead), propagate to the right fronto-central area, and change to irregular spike waves. Finally, waves suddenly terminate (arrow). (C) After administrations of tocilizumab on day 246 of admission, EEG reveals diffuse, low-amplitude slow waves, and no epileptic discharges. TC = time constant; HF = high-frequency filter.
Previous studies have mentioned the relationship between epilepsy and IL-6 (Vezzani et al., 2011;Peltola et al., 2000). Repetitive or prolonged epileptic seizures lead to the production of proinflammatory cytokines, including IL-6, by activating microglia, astrocytes, and neurons. The inflammatory cascades activated in this manner induce disruption of the blood-brain barrier and further inflammation, thereby sustaining seizure activity (Vezzani et al., 2011). Previous studies have reported that concentrations of IL-6 in CSF and plasma were significantly increased after convulsive seizures (Peltola et al., 2000), and that anti-IL-6 treatment reduced the development of seizures in rat models of epilepsy (Leo et al., 2020). Intrathecal overproduction of IL-6 has also been described in pediatric febrile infection-related epilepsy syndrome, which corresponds to adult NORSE (Sakuma et al., 2015), suggesting that anti-IL-6 treatment might be effective against NORSE. Jun et al. reported 7 NORSE cases (6 cryptogenic cases, 1 case of anti-NMDAR encephalitis) treated with 1 or 2 doses of tocilizumab, with SE resolving immediately in 6 patients (Jun et al., 2018). Since our patient revealed negative results for already-known anti-neuronal autoantibodies and indirect immunofluorescence, and presented with elevated levels of IL-6 in CSF, the pathogenesis in this patient might not have been induced by autoantibodies, but instead by IL-6-mediated inflammatory responses. We therefore used tocilizumab in the present case. According to the immunotherapy regimens of the above case series (Jun et al., 2018), we started to administer tocilizumab at a dosage of 4 mg/kg for 2 cycles in 1-week intervals, then a third time at a dosage of 8 mg/kg almost one month later, finally achieving total control of SE. Thus, we presume anti-IL-6 treatment may prove effective for stopping SE in some C-NORSE cases, particularly those presenting with elevated levels of IL-6 in CSF.
The above case series suggest that early treatment with tocilizumab might prove more effective for better functional outcomes, since patients with good or fair outcomes (mRS score ≤ 3) in this study exhibited a relatively short duration of SE (median, 11 days) compared to other patients (median, 42 days) (Jun et al., 2018). Compared to the case series above, our case experienced a much longer duration of SE (140 days), needed more doses of tocilizumab (3 doses), and experienced a poor outcome (mRS score 5). One reason our case experienced a poor outcome may be the irreversible damage induced in the cerebral cortex by very long-lasting SE. If we had performed anti-IL-6 treatment earlier before the onset of irreversible neurological disorders, our patient might have achieved better functional outcomes. More cases and studies clearly need to be accumulated to confirm this hypothesis.

Funding
This research did not receive any specific grants from funding agencies in the public, commercial, or not-for-profit sectors.

Declaration of Competing Interest
None.