Léa Cosme

Léa Cosme

Auto-generated profile, based on publicly available information

Léa Cosme

Epilepsy Unit and Department of Clinical Neurophysiology AP‐HP, Hôpital de la Pitié‐Salpêtrière, Sorbonne Université, ERN‐EpiCARE Paris France

Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
16 February 2026 in Epilepsia
Verified
Journal Article
Open access
AHAurélie Hanin
Aurélie Hanin
CMClémence Marois
Clémence Marois
MGMartin Guillemaud
Martin Guillemaud
MCMario Chavez
Mario Chavez
LCLéa Cosme
Léa Cosme
VNVincent Navarro
Vincent Navarro
Objective: Cryptogenic new onset refractory status epilepticus (cNORSE) carries high risks of long‐term disability and post‐NORSE epilepsy, but mechanisms remain unclear. We aimed to assess the predictive value of inflammatory and brain injury biomarkers and determine whether immune disturbances persist in the chronic phase. Methods: We enrolled 93 cNORSE patients from the Pitié‐Salpêtrière Hospital and the Yale NORSE/FIRES biorepository (2013–2025). Serum and cerebrospinal fluid (CSF) samples were collected during status epilepticus (SE), with outcomes assessed 6–12 months after resolution. To investigate post‐cNORSE epilepsy, we compared 39 post‐cNORSE patients (25 with paired acute samples) to 40 patients with temporal lobe epilepsy due to hippocampal sclerosis (TLE‐HS) and 20 with chronic immune‐mediated encephalitis. Results: During cNORSE, elevated innate cytokines (serum CXCL8, CCL2; CSF IL‐6, CXCL8, CCL2, MIP‐1α, G‐CSF) and brain injury biomarkers (serum and CSF neurofilament light chain [NfL], CSF neuron‐specific enolase) correlated with worse functional outcomes. Multivariate models demonstrated that adding serum NfL to cytokines improved poor outcome prediction (area under the curve = .75). In contrast, no acute biomarker predicted post‐cNORSE epilepsy, which was instead associated with prolonged SE, magnetic resonance imaging abnormalities, and the need for more intensive treatment. In paired analyses, most serum cytokines normalized during the chronic phase, particularly IL‐6, IL‐10, and IL‐1β, although new adaptive immune disturbances (IL‐17A, IL‐12p70, TNFα) appeared in 20% of patients. No chronic elevations of innate cytokines were observed in post‐cNORSE patients. Conversely, elevated age‐adjusted NfL levels were more frequent in post‐cNORSE epilepsy (64%) than encephalitis (45%) and TLE‐HS (20%), ( p < .001), with elevated NfL levels correlating with poor functional outcomes ( p = .019). Significance: Innate immune activation is a hallmark of acute cNORSE but largely resolves in the chronic phase, arguing against persistent innate inflammation as the driver of post‐cNORSE epilepsy. In contrast, persistently elevated NfL levels suggest ongoing axonal injury, potentially contributing to poor outcomes. Integrating inflammatory and neuroaxonal injury biomarkers may improve risk stratification and guide long‐term management.

5 November 2026 in Epilepsia
Verified
Journal Article
Open access
AHAurélie Hanin
Aurélie Hanin
CMClémence Marois
Clémence Marois
LCLéa Cosme
Léa Cosme
QCQuentin Calonge
Quentin Calonge
ZHZineb Hayatou
Zineb Hayatou
VNVincent Navarro
Vincent Navarro
Objective: To characterize serum and cerebrospinal fluid (CSF) cytokine dynamics according to immunotherapy exposure in cryptogenic new‐onset refractory status epilepticus (cNORSE). Methods: We analyzed cytokine profiles in 94 serum and 42 CSF samples from 29 and 14 patients, respectively, with cNORSE. Nine contributed both sample types, resulting in 34 patients. Samples were categorized according to immunotherapy exposure as no immunotherapy (NT), first‐line (FL), and second‐line (SL), and stratified by innate‐ or adaptive‐directed therapies. SL represented cumulative exposure, as all SL samples followed initiation of both FL and SL therapies. Linear mixed‐effects models accounted for repeated measures and time from cNORSE onset. A systematic literature review included 27 studies and 63 patients. Results: Serum cytokine levels were generally lower during FL ( n = 24) than NT ( n = 9), whereas most cytokines were higher during SL ( n = 61) than FL. CXCL8 (C‐X‐C motif ligand 8)/IL‐8 and CCL2 (C‐C motif chemokine ligand 2) were higher during SL than FL, with CCL2 remaining significant after multiple‐testing correction. No serum differences remained significant when analyses were restricted to samples collected 7–30 days after onset. CSF levels were generally lower during FL ( n = 10) and SL ( n = 25) than NT ( n = 7). Innate‐directed therapies were associated with higher serum IL‐6, CXCL8/IL‐8, and CCL2 and CSF IL‐4, whereas adaptive‐directed therapies were associated with lower CSF MIP‐1α (Macrophage Inflammatory Protein‐1 alpha). The association with higher serum IL‐6 was no longer observed after excluding tocilizumab‐treated patients. No temporal association after SL initiation or status epilepticus resolution remained significant after multiple‐testing correction. The literature review showed more frequent reductions in pro‐inflammatory cytokines in CSF than serum following immunotherapy initiation. Significance: Cytokine profiles in cNORSE differ by treatment exposure and compartment. Persistent serum inflammation during SL may reflect disease severity, systemic inflammation, treatment effects, or their combination, whereas CSF showed more consistent reductions after immunotherapy exposure. These findings highlight the importance of treatment timing and compartment when interpreting cytokine measurements and support prospective studies integrating baseline immune phenotypes and standardized clinical outcomes.