Clinical manifestations

Presentation and course

NORSE and FIRES follow a biphasic course, with a prodromal phase preceding the onset of status epilepticus and encephalopathy by up to 2 weeks. Status epilepticus often has a subacute onset over a few hours to days. The prodromal phase, which is characterized by nonspecific mild illness with gastrointestinal, upper respiratory, or flu-like symptoms, often precedes the onset of seizures (21; 79). Headaches and hallucinations are also often reported (44). In children, a preceding febrile episode, which defines FIRES, is almost always reported (90%) (79; 96). In adults, fever is reported in only 30% to 50% of cases (21; 60), although it seems to be more frequent in cryptogenic cases (60% to 90%) (21; 43; 11). This difference could be explained by the greater propensity of fever to appear or be noticed in children than in adults (44).

Seizures typically appear between 24 hours to 2 weeks after the prodromal phase, often with a short free interval. Cognitive and behavioral symptoms, including hallucinations and memory difficulties, can occur before or at the same time as seizures. Seizures are initially brief and infrequent, increasing within a few hours to days in frequency (up to hundreds per day) and severity, together with progressive impairment of consciousness. They quickly evolve into status epilepticus that persists or recurs despite administration of anesthetics in continuous infusion, corresponding to (super) refractory status epilepticus requiring intensive care. The most frequent seizure type is focal seizure with or without impairment of awareness and with secondary bilateralization and multifocal onset (52; 40; 21). A temporal-perisylvian semiology is often seen (66).

Prognosis and complications

The episode of status epilepticus can last several days or weeks. The majority of FIRES and NORSE cases evolve to super-refractory status epilepticus, which is defined as status epilepticus persisting for at least 24 hours, or recurring while on appropriate anesthetic treatment, or reoccurring after withdrawal of anesthesia. In adults, the median duration of status epilepticus is 17 days and of ICU stay is 15 to 30 days (21; 60; 11). The mortality rate in adult NORSE is 16% to 27% (10; 19; 21; 48; 33). Long-term outcome is often poor, with half to two-thirds of the survivors developing cognitive impairment, behavioral disorder, and functional disability (05; 53; 65; 75; 40; 08; 21; 67; 87; 79; 11). Subsequent drug-resistant epilepsy develops without a latent period in greater than 60% survivors (85; 11; 33). A diffuse brain atrophy appears in one to two thirds of cases in both adults and children (53; 40; 08; 43). Diffuse brain atrophy appears in up to two third of cases both adults and children (53; 40; 08; 43; 33).

In children, the range of status epilepticus duration varies across the literature. A retrospective case series of 92 children with NORSE, including 90% with FIRES, showed a median ICU stay of 5 days, with a status epilepticus duration of 8 days (96). The mortality was 23%, similar to adults, with 40% of survivors having poor functional outcome. Previous studies on pediatric FIRES tended to report longer status epilepticus duration (52) but lower mortality rates (12%) and worse outcome on survivors, with refractory residual epilepsy in many of them and 50% to 60% with cognitive disability (half of them severe, including vegetative state) (52; 79). Executive dysfunction is most prominent (63; 52).

Super-refractory status epilepticus (96), diffuse cortical edema or multifocal abnormalities observed on MRI (96), nonconvulsive status epilepticus (96), young age (52), and duration of barbiturate coma (52) were the factors associated with worse outcome. It is unclear if this association indicates an independent effect of anesthesia or if it reflects the intrinsic severity of the underlying disease that required more intensive treatment.

Biological basis

Etiology and pathogenesis

NORSE and FIRES cover a heterogeneous group of disorders. In approximately 30% of cases, a definite etiology is ultimately found. Despite an extensive workup, greater than 75% (21; 60) and children (52; 96) with NORSE and FIRES remain of unknown etiology (also known as cryptogenic NORSE or c-NORSE). These cases might share a common pathogenesis.

Overall, clinical features are similar in c-NORSE and autoimmune cases. However, the status epilepticus tends to be more refractory with response to first-line immunotherapy, far more disappointing than expected in autoimmune encephalitis, mainly in children. They often have a worse outcome, with residual refractory epilepsy in more than 90% of survivors and cognitive sequelae in more than 80% of FIRES (63; 75; 52; 40; 79). Only 4% of c-NORSE patients return to baseline, and 40% achieve favorable functional outcome (11).

Immunology. Even if the exact pathogenesis is still not fully understood, several studies suggest that FIRES and NORSE of unknown etiology might be due to post-infectious genetically determined autoinflammatory mechanisms (89), involving cytokines and chemokines of the Th1-related immunity and inflammation pathways (76; 46; 45; 09; 51; 41). A study showed specific immune signatures of NORSE and FIRES of unknown etiology, compared to patients with refractory status epilepticus and NORSE with a known etiology, characterized by increased serum levels of C-X-C motif ligand (CXCL)8, C-C motif ligand (CCL)2, macrophage inflammatory protein (MIP)1a, and interleukin (IL)10 and IL4, and increased CSF levels of IL1beta (31). The level of several of these cytokines correlated with daily ictal burden, severity, and worse outcome (31). Cytokine levels are also more elevated in children with FIRES than in adults (44). These cytokines are possibly produced by hyperinflammatory macrophages that exhibit abnormal response to exogenous triggers, such as viral or bacterial antigens (39). Of note, the abnormal cytokine profile found in NORSE and FIRES of unknown etiology differs from the one observed in antibody-mediated autoimmune encephalitis, which results from a response of the adaptive immune system with elevations in TNF-alpha, IL-2, IL-12, IL-4, and IL-10 and may explain the greater response to first-line immunotherapies in these conditions (45). Several of the cytokines elevated in cryptogenic cases are known to increase neuronal excitability in vivo, either directly or indirectly, through a variety of mechanisms. In turn, seizures themselves can further activate the immune system through the increase of cytokine levels, by activated glia and blood-brain barrier leakage, thus, creating a self-perpetuating cycle of inflammation and hyperexcitability. Studies are ongoing to further investigate these mechanisms at the single-cell level (33).

Genetics. The autoinflammatory mechanism underlying the pathogenesis of NORSE and FIRES of unknown etiology is driven in part by genetic predisposition. Genetic studies have revealed the presence of polymorphisms in the IL1RN gene, which codes for a soluble antagonist of the IL-1beta receptor (sILRA) that modulates the IL-1beta pathway in patients with cryptogenic FIRES. The resulting sIL1RA protein appears to be functionally deficient, leading to an increase of IL-1beta production (50; 09). Another study demonstrated that children with cryptogenic FIRES have defective toll-like receptor (TLR) signaling despite normal TLR expression on peripheral monocytes, as well as decreased naive T and T regulatory cells and weakened phagocytosis (41). This could not only increase the hosts’ susceptibility to viral infections but also decrease their ability to eradicate pathogens. In turn, prolonged exposure to viruses and accumulated cell and pathogens debris, in conjunction with decreased T regulatory cells, could promote autoimmunity and autoinflammation. Thus, these studies imply a genetic predisposition to cryptogenic NORSE and FIRES (74; 09) and provide the first hint towards the mechanisms that lead to increased levels of anti-inflammatory cytokines in patients with NORSE or FIRES (09; 51). Altogether, cryptogenic NORSE and FIRES would then result from an aberrant post-infectious neuroinflammatory response.

In line with this hypothesis, several case reports and series have described spectacular clinical and electrographic responses to therapies targeting the innate inflammatory pathway like anakinra, a recombinant IL-1 receptor antagonist, or tocilizumab, an IL-6 receptor antagonist (46; 45; 14; 72; 93; 55; 83; 98; 01). Use and effect of those treatments are detailed in the therapy section below.

Epidemiology

NORSE and FIRES can affect people of all ages, most frequently young adults and school-aged children (older than 2 years) or adults over 60 years old (mainly afebrile NORSE) (29). Adult patients with FIRES seem to be younger than patients with nonfebrile NORSE (44). The incidence is unknown, but it can be estimated that NORSE represents up to 20% of cases of refractory status epilepticus (20; 20). There is a female predominance in most adult series (10; 21; 48; 85), whereas in children, boys tend to be more frequently affected than girls (52; 96). A seasonal variation in the incidence of NORSE has been found in some (06) but not all studies (27).

Differential diagnosis

Associated or underlying disorders

The possible causes of NORSE and FIRES are many (56). They can be classified into four categories: (1) inflammatory and autoimmune encephalitis, (2) uncommon infectious encephalitis, (3) genetic disorders, and (4) toxic disorders (20).

The etiologies vary with age. For instance, a paraneoplastic etiology is more frequent in adults (18%) than in children, whereas uncommon infection is more frequently reported in children than adults (20% vs. 10%) (21; 26; 42; 96; 56). Metabolic and genetic etiologies, including mitochondrial diseases, are also more frequent in children, but they can rarely present with NORSE in adulthood.

In young adults, the most frequently identified cause is autoimmune encephalitis (19% to 37% of all NORSE cases) (21; 60), including sporadic and paraneoplastic cases. Antibodies directed against neuronal cell surface antigens are the most involved. The most frequently identified antibodies target the N-methyl-D-aspartate (NMDA) receptor (12% to 15%). Antibodies against leucine-rich glioma inactivated 1 (LGI1) (21; 56), gamma-aminobutyric acid (GABA)bR, and GABAaR are also found. NORSE associated with onconeuronal antibodies also exist (21). Antibodies are exceptionally found in pediatric cohorts (3%) (79; 96), with anti-NMDA, contactin-associated protein-like 2 (Caspr-2) (96), or GABAaR (56). One study found anti-glutamate receptor (GluR) epsilon 2 antibodies in the CSF of a few FIRES cases, but their role and significance remain unknown (75).

Some clinical features can suggest a specific underlying etiology and are important to recognize. For instance, paraneoplastic limbic encephalitis is characterized by cognitive impairment, behavioral changes, sleep disturbances, and seizures with eventual progression to status epilepticus (03; 20). Anti-NMDA receptor encephalitis is the first cause of overall autoimmune encephalitis (13) and often starts with febrile illness. Patients then develop psychiatric symptoms manifesting in children as behavioral disturbances and tantrums. Children are more likely to present movement disorders, seizures, and status epilepticus than adults (18; 03; 70). Language regression, hyperactivity, and irritability are often seen, followed by progression to decreased responsiveness and severe catatonic stage, with typical orolingual dyskinesia and autonomic failure (70). The EEG shows a specific pattern of “extreme delta brushes” in 50% of cases (77; 30). Encephalitis with antibodies targeting the voltage-gated potassium channel (VGKC) complex (LGI1 or, more rarely, CASPR2) are associated with limbic encephalitis and a syndrome of inappropriate secretion of antidiuretic hormone (SIADH). Pathognomonic faciobrachial dystonic seizures can occur in anti-LGI1 encephalitis.

Although genetic investigations are usually disappointing (02; 36), a few mutations have occasionally been identified in sporadic NORSE and FIRES cases (42; 96; 91; 94; 23), including point mutations in the KCNT1 (potassium sodium-activated channel subfamily T member 1), CACNA1A (calcium voltage-gated channel subunit alpha1A), POLG1 (DNA polymerase gamma subunit 1), PCDH19 (protocadherin 19), cathepsin D, CEP290 (centrosomal protein of 290 kDa), POU1F1 (POU class 1 homeobox 1), FADD (FAS-associated protein with death domain), and DNM1L (dynamin 1 Like) genes and a chromosomal 9p24.3 chromosomal duplication.

Diagnostic workup

Both brain MRI and CSF analyses are the first procedures required when a diagnosis of NORSE or FIRES is suspected in order to exclude other causes of acute symptomatic status epilepticus, like stroke or infectious encephalitis. The second step aims to identify eventual causes of NORSE, mainly using antibody panels and metabolic and genetic analyses.

CSF. Half to two thirds of cases present with mild CSF pleocytosis (usually less than 10 cells/µl) and slightly increased CSF protein level (52; 21). By definition, first-line microbiological studies should be negative.

MRI. In the acute phase, about 30% to 50% of pediatric (79; 96) or adult (21; 60) NORSE cases have an abnormal initial MRI. The most frequent finding is with T2/FLAIR hypersignal located in limbic or neocortical areas, often bilaterally and symmetric (43). In children, diffuse cerebral edema is often reported (12; 96). Basal ganglia (63), claustrum (62; 61), and peri-insular (08) involvement have also been reported. Two studies have highlighted the presence of peculiar, and possibly specific, transient bilateral claustral T2/FLAIR changes in adult and children with NORSE or FIRES of unknown etiology (62; 61). It is unclear if these findings are specific to NORSE or if they are the consequence of prolonged ictal activity. A normal MRI is possibly associated with better outcome in children (79; 96).

EEG. The EEG reveals sporadic or periodic epileptiform discharges, which can be lateralized, bilateral independent, or multifocal, often involving the temporal and frontal regions. Generalized discharges have also been reported, mainly in children (50%) (21; 79). A retrospective study identified three EEG findings in FIRES: (1) beta-delta complexes (20%) (60), (2) seizure onset with prolonged focal fast activity, followed by the gradual appearance of well-formed rhythmic spike or spike-and-wave complexes, and (3) shifting ictal activity (17). Those findings should be confirmed by further studies. Continuous EEG monitoring is indicated as most seizures in refractory and super-refractory status epilepticus are either purely electrographic or clinically too subtle to allow clinical monitoring.

C-NORSE score. No validated test is available to identify NORSE and FIRES cases of unknown etiology (in the first hours of status epilepticus onset). Therefore, the diagnosis is established only after a few days, which is the time required to obtain the results of autoimmune screening. As data suggest quickly treating patients more aggressively, an earlier diagnosis is required (94). A cryptogenic NORSE score has been proposed for this purpose (97). This score includes the following features: presence of prodromal high fever of unknown origin before the onset of status epilepticus, absence of prodromal behavioral or memory alterations before onset of status epilepticus, absence of sustained orofacial-limb dyskinesias despite a profoundly decreased level of consciousness, and symmetric brain magnetic resonance imaging abnormalities. Each feature is worth 1 point, and a high score suggests crypotgenic NORSE. The sensitivity and specificity of a high score for predicting C-NORSE are 94% and 100%, respectively, potentially easing the decision to initiate immunotherapy targeting the innate immune system as early as 48 hours from the onset of status epilepticus.

Autoimmune, metabolic, infectious, and genetic testing. Because of the rarity of the syndrome, work-up strategies markedly vary between institutions and clinicians (56). Whereas infectious and autoimmune testing are often performed, genetic testing is performed in only a minority of cases.

Delphi methodology-based consensus recommendations for standardized work-up are now available and should improve diagnostic yield (94). According to these guidelines, an initial standardized diagnostic workup is recommended for all NORSE and FIRES cases, regardless of age and the presence of a prodromal febrile episode (94). This initial work-up should mostly aim to identify autoimmune and infectious encephalitis, and complete autoimmune workup is essential in any NORSE case given that autoimmune encephalitis is the most frequent identified cause of NORSE in adults.

Based on the results of this initial work-up, further investigations need to be tailored according to the patient's characteristics (80). For instance, young children should more promptly undergo extensive metabolic and genetic investigations as the yield of these investigations is higher in this age group. Immunosuppression and known exposure to potential pathogen vectors should heighten the suspicion of uncommon infectious causes (94). A malignancy screen, including whole body PET, should be considered in older adults. However, early testing for autoimmune antibodies should nearly always be done in all NORSE cases as results will impact the treatment choice.

Neuropathology. Very little data are available, and findings are variable. The most frequent findings include neuronal loss, reactive gliosis, microglial activation, meningeal inflammation, and perivascular T-cell infiltration (31).

Management

There is no randomized controlled trial for the treatment of NORSE patients, and most institutions do not have standardized protocols to manage them. Recommendations have been published, relying on expert opinion through a Delphi method, in an effort to share experiences and create a unified standardized approach (94). Second-line immune therapies seem to be increasingly used (33).

The initial treatment of NORSE and FIRES in the first 24 to 48 hours does not differ from the treatment of status epilepticus in general. Early status epilepticus requires prompt treatment, usually with a parenteral benzodiazepine and, if unsuccessful, with an IV load of a nonsedating antiseizure medication. When refractoriness is established, patients should ideally be admitted in a tertiary center with continuous EEG available and multidisciplinary expertise. The course of the treatment is then guided by continuous EEG to identify electrographic seizures, as it is recommended for all patients with refractory status epilepticus (94). Treatment of NORSE with antiseizure medications is often disappointing. At least 75% of patients require anesthetics in continuous infusion and often prolonged burst-suppression coma to stop the seizures. Status often resumes once the anesthetics are weaned off (52; 89). Of note, some FIRES experts recommend avoiding anesthetics in continuous infusion as it has been associated with poor outcome in several studies; however, those data could reflect allocation bias, as the more severe cases likely received more intensive sedation (49).

Once the situation is recognized as NORSE, usually within 48 to 72 hours from the onset of status epilepticus, early immunotherapy is recommended. In cryptogenic and proven or suspected autoimmune NORSE, corticosteroids are the most common first-line treatment, often combined with IVIG. But IVIG alone can be used if an infectious etiology is strongly suspected and has not yet been excluded (94). If the patient does not improve, experts suggest initiating the ketogenic diet, mostly in children, or second-line immunotherapies within 1 week of the onset of status epilepticus once infectious etiologies have formally been ruled out. Response to first-line immune therapies in antibody-mediated cases is often better than in cryptogenic cases, and rituximab can be administered as a second-line therapy.

In cryptogenic NORSE and FIRES, response to first-line therapies and rituximab is often disappointing, and other second-line immunotherapies should be initiated (75; 52; 88; 84).

Given the suspected underlying role of the innate immune system, immunotherapy blocking proinflammatory innate cytokines, such as anakinra, tocilizumab, intrathecal dexamethasone, and the ketogenic diet have been tried, with promising results (45; 14; 38; 33).

Anakinra. Overall, approximately 70 patients, mostly children with cryptogenic FIRES, have been reported in the literature. A retrospective cohort of 25 children showed that earlier initiation of anakinra was associated with shorter duration of hospital stay, and 11 out of 15 children had a more than 50% decrease in seizure burden after 1 week (55). This treatment was also associated with normalization of IL-8 and IL-6 levels, with an improvement of epilepsy in several cases (46), suggesting that inflammation biomarkers could be used to assess treatment efficacy (54). Despite the occurrence of severe adverse reactions in two thirds of patients, mainly cytopenia and infections, anakinra was rarely discontinued (55). A risk of DRESS syndrome also exists (71).

Tocilizumab. Overall, approximately 35 patients, mostly adults with cryptogenic FIRES or nonfebrile NORSE, have been reported in the literature. Tocilizumab was also reported to be successful in six of seven adult patients with NORSE, with status epilepticus resolved after one or two doses and an interval of 3 days from the onset of treatment. Better outcome was associated with earlier treatment. Adverse events occurred in five patients: leukopenia in three cases, sepsis, and pneumonia (45). Several case reports also highlighted the efficacy of this treatment, both in children and adults (07; 15; 92), and even in FIRES cases refractory to anakinra (83; 01). The experts recommend starting either anakinra or tocilizumab within the first week of onset of status epilepticus. Measuring serum and CSF IL-6 levels in the first few days could be useful to monitor the response to treatment (24).

Current evidence does not support the use of one over the other molecule. Tocilizumab is a larger molecule than anakinra and can be less effective in penetrating the blood brain barrier (59). Consequently, the peripheral concentration of IL-6 could increase and induce a paradoxical neurologic deterioration (28). The duration of treatment is not yet well defined, and further studies are required to provide evidence-informed recommendations for both drugs, which are currently not FDA or EMA approved for this indication. If well tolerated, immunomodulation treatments should be continued for 3 months, especially if cytokines were increased in the acute phase (94).

Intrathecal dexamethasone. The use of intrathecal dexamethasone (IT-DEX) has been reported in a total of 11 patients, including nine children. In the initial report of 6 children with FIRES IT DEX was administered at least four times with an interval of 1 to 6 days. This allowed reduction of 50% of seizure burden in three of them with improvement of EEG background (38). The main advantage of IT-DEX compared to anakinra and tocilizumab seems to be the absence of severe adverse event.

Ketogenic diet. The use of the ketogenic diet in NORSE and FIRES was reviewed (64). It was first reported to be successful in seven of eight children with FIRES, with improvement noted after 1 to 4 days of ketonuria (65). This was further confirmed by a case series in which all seven children improved by ketogenic diet, which was sometimes administered by intravenous line (68). The efficacy of the ketogenic diet in super-refractory status epilepticus in adults, including possible NORSE cases, has also been suggested (86). In FIRES, ketogenic diet should be introduced within 2 to 5 days of the onset of status epilepticus (49). There have been several reports on the efficacy and safety of the ketogenic diet in this setting; a total of 72 cases were included in a systematic review of the ketogenic diet in pediatric super-refractory status epilepticus (78). Overall, ketosis was achieved in 96% of cases, within a mean time of 3.4 days, and status epilepticus resolved in 60% of cases, within a mean of 6.3 days after the initiation of ketogenic diet. Adverse events occurred in a third of children and were mostly mild gastrointestinal side effects. Hypoglycemia, metabolic acidosis, weight loss, and nephrolithiasis occurred in less than 5% of cases (81). The ketogenic diet induces an anti-inflammatory process through the inhibition of IL-1beta and, therefore, could act in synergy with anakinra (69).

Other less common treatments have been proposed. In a few cases series on refractory status epilepticus, hypothermia at 33°C was found to be effective in controlling seizures in adults or children (57), but only rare case reports refer to NORSE or FIRES (58). The risk of complication is substantial, manly aspiration pneumonia. Thus far, there are no sufficient data to recommend this therapeutic option. Cannabidiol is a potential alternative therapy in epilepsy (04) and has been shown to improve seizure frequency and duration in six out of seven children with FIRES, but mainly in the chronic phase (25). Electroconvulsive therapy and neurostimulation with vagal nerve or deep brain stimulation have also been anecdotally used (82).

All these therapeutic options need to be validated by large and prospective studies. A suggested algorithm for the use of these therapies is summarized, with their suggested dosing.

Outcomes

Even fewer data are available concerning epilepsy in the post-acute phase of NORSE and FIRES. Combinations of multiple antiseizure medications are usually administrated given the residual refractory epilepsy in those patients. No specific medication can be preferentially recommended. Most treatment initiated in the acute phase with an effective response can be continued in the post-acute phase (94), but in most cases, long-term immunotherapies are limited by the risk and occurrence of side effects. The doses can sometimes be adapted. For instance, maintenance steroids should be avoided, but intermittent steroid pulses can be considered (47). Long-term immunomodulation for more than 3 months is not encouraged, but re-challenges are sometimes performed in cases of worsening on withdrawal. One small study suggests a positive effect of treatment targeting IL-6 and IL-8 in the chronic phase of the disease, but this needs to be confirmed by larger trials (01). The mechanism of cognitive alteration in NORSE is likely related to the severity and duration of status epilepticus. Persistent inflammation and structural change may also play a role that could explain those successful responses of immunotherapy in the chronic phase (85).

Data on neurostimulation, presurgical work-up, and resective surgery are lacking (94). Psychological and behavioral counseling and family support are often needed.

As explained in the prognosis section, NORSE and FIRES are associated with a high rate of mortality, and most survivors have moderate to severe cognitive disabilities. Therefore, it is recommended to periodically repeat a neuropsychological evaluation in the follow-up care of those patients. Many will need to undertake an intensive program of cognitive rehabilitation (94). Given the retrospective and observational design of most studies, with limited number of patients and heterogeneous protocols, the efficacy and safety of the different treatments cannot be clearly assessed nor compared. Clinical trials with standardized approach are needed, using uniformed scales and tools to measure seizure burden but also functional outcomes, including in the chronic phase.