NORSE, FIRES, and a Polygenic Trickle of Autoimmunity

There are acute presentations of epilepsy that are so dramatic that clinicians remember the specific scenarios’ years, if not decades later-seizures that suddenly start without a reason and simply do not stop despite state-of-the-art treatment, a condition referred to as new-onset refractory status epilepticus (RSE). Despite their acuity, the intensity of care, and a mortality of up to 30%,¹ presentations like this are rare, which has made meaningful research into the underlying causes and treatments difficult. However, over the last 2 decades, a community of clinicians, researchers, and advocates have come together to advance our understanding of these enigmatic conditions. First and foremost, by giving them a name and working definition.

New-onset refractory status epilepticus (NORSE) is typically diagnosed if an identifiable structural, toxic, or metabolic cause cannot be recognized.² Febrile infection-related epilepsy syndrome (FIRES) is a subset of NORSE defined by refractory status epilepticus (RSE) that is preceded by a febrile illness.³ Despite differences in definitions, the diagnostic label NORSE is typically used in adults, whereas FIRES is typically used in a pediatric setting. Very rare presentations of NORSE are caused by genetic disorders, such as POLG, PCDH19, RANBP2, or CACNA1A,⁴ or by autoimmune or infectious encephalitis. However, the majority remains unexplained, and the term cryptogenic NORSE (cNORSE) is used to clarify the lack of identifiable causes. The situation in FIRES is not much different. Despite an average 15% yield of genetic causes in severe epilepsies without identifiable etiologies, the diagnostic rate in FIRES is zero.⁵ In summary, despite a tangible proximity to both genetic epilepsies and inflammatory and autoimmune condition, cNORSE and FIRES remain unexplained.

In their recent publication,⁶ Jang and collaborators took a different approach to assess a potential genetic contribution to cNORSE. In a carefully phenotyped sample of 30 individuals, the authors performed whole-genome sequencing and analyzed polygenic risk scores (PRSs) for various conditions. In contrast to diagnostic exome or genome sequencing that aims at the identification of causative, rare variants, PRSs accumulate the effect of common genetic variations, so-called single nucleotide polymorphisms (SNPs). The contribution of common genetic variants is typically assessed through genome-wide association studies, large cohort studies with thousands of individuals aiming at identifying SNPs that are more common in cases than in controls. After these studies have been performed, associated SNPs and their relative contribution can be combined into a PRS. These PRSs can then be used to assess genomic overlap between conditions even in smaller samples.

PRSs are becoming more important in the epilepsy, particularly in conditions where a single monogenic cause cannot be identified.⁷ Recent studies have shown that PRS can provide valuable insights to better delineate and understand about an etiology.⁸ Here, Jang and collaborators used this method to assess the overlap with 40 conditions for which PRSs were available. Although none of these conditions showed an overlap significant after Bonferroni correction, the ranking of conditions with marginal significance was perplexing. Rather than an expected overlap with epilepsies, cNORSE showed the strongest overlap with autoimmune conditions, such as ulcerative colitis, systemic lupus erythematosus (SLE), or Hashimoto's disease. The authors then performed several additional analyses and found variants associated with cNORSE were enriched in the central nervous system (CNS), particularly the hippocampus, as well as in lymphocytes. In brief, an autoimmune signature with a predilection for the CNS.

Before diving into the potential implications, it is important to highlight a major limitation of the study by Jang and collaborators.⁶ The sample size was small and none of the findings were significant after multiple testing. Rather than a stand-alone result, the study by Jang and collaborators⁶ is a call to action for the community to independently replicate these findings and put the genomic overlap with autoimmune conditions on solid footing. NORSE and FIRES are rare conditions. However, the rarity of a condition does not affect the statistical threshold needed to achieve significance—independent replication (or refutation) is urgently needed. Otherwise, such findings may become a self-fulfilling prophecy, a narrative that adds to the long-standing consideration of autoimmunity in NORSE and FIRES rather than speaking for itself. Additionally, although these findings will most likely motivate a new era of research into autoimmune causes of NORSE and FIRES, it is important to keep in mind the unsuccessful efforts that have already been spent trying to identify such causes, including the lack of an HLA signature in FIRES.⁹ In addition, PRSs are not iron-clad—they depend on many assumptions and arbitrary cutoffs, further emphasizing the need for replication rather than taking the findings by Jang et al⁶ as ground truth.

Putting this limitation aside, let us speculate for a moment what these findings might imply. If confirmed, cNORSE could represent an unusual condition—a potential autoimmune condition identified not through autoantibodies but through a genomic relatedness to other autoimmune disorders. Although compelling given the lack of alternative explanations, there is little precedence in genomic research how such indirect evidence can be translated into clinical practice. The acute treatment of NORSE and FIRES often occurs with immense urgency, given the intractability of seizures. The findings by Jang and collaborators may face clinicians with the choice of using immunosuppressive therapies more broadly. In fact, a recent review of second-line immunotherapy in NORSE and FIRES suggested a potential utility for anakinra, tocilizumab, and intrathecal dexamethasone.¹⁰ However, in the authors’ own words, it is important to keep in mind that genomic overlap does not imply causation.

Despite these precautions, Jang and collaborators should be congratulated for bringing one of the unsolved problems of epilepsy genetics to the forefront—the elusive genetic basis of NORSE and FIRES. Their study represents a watershed moment for research into these rare conditions, leading us into a new direction. Hopefully, their finding provides new momentum to a field that has long been devoid of tangible hypotheses to explore.