Abstracts

Rationale: Interleukin 1 receptor antagonist (anakinra) has been used increasingly for the treatment of new onset refractory status epilepticus (NORSE) and febrile-infection related epilepsy syndrome (FIRES). Presently, there are no known biomarkers for anakinra response or early indicators of neurological outcome. We evaluated whether features of continuous electroencephalography (cEEG) could serve as indicators of either anakinra response or neurological outcome in children with NORSE/FIRES who were treated with anakinra.

Methods: We analyzed a 24-hour cEEG epoch recorded from 4 anakinra responders (3 male, 1 female) and 2 non-responders (2 males) at the following epochs: 24 hours prior to, 1 week, 2 weeks, and 3 weeks following anakinra initiation. We analyzed cEEG background, frequency of seizures and epileptiform discharges, presence of generalized or lateralized periodic discharges (GPDs/LPDs) and highly epileptiform bursts (HEBs). We quantified seizure frequency as number of seizures/24 hrs. We categorized the prevalence of the epileptiform discharges as 0%, < 1%, 1 - 10%, 10 - 50%, 50 - 90%, and > 90%. We also assessed outcome using pediatric cerebral performance category at 1 year from seizure onset or the latest available time point. The outcome was dichotomized as good (normal/mild disability) or bad (moderate disability to vegetative state). The outcomes of interest were cEEG features associated with anakinra response and those associated with good outcome. Data were analyzed using Chi square or Fisher exact test.

Results: At 24 hours prior to anakinra initiation, 2 responders were on midazolam, 1 on pentobarbital and 1 on both. All non-responders were on both midazolam and pentobarbital infusions. Two anakinra responders had mild disability; 2 had moderate disability. All non-responders had severe disability. One responder had seizures prior to anakinra, which resolved following anakinra. One non-responder had persistent seizures before and after anakinra initiation. There were no differences in the frequency of epileptiform discharges following anakinra between responders and non-responders. Both responders and non-responders continued to have GPDs/LPDs. Prior to anakinra initiation, 2 responders and 1 non-responder had HEBs. Following anakinra initiation, responders had fewer cEEGs with HEBs as compared with non-responders (responder: 2 of 12 vs. non-responder: 6 of 6, p = 0.002). At 2 and 3 weeks following anakinra, HEBs were no longer observed in the cEEGs of the responders (responder: 0 of 8 vs. non-responders: 4 of 4, p = 0.002). Compared to children with bad outcome (n = 4), children with good outcome (n = 2) had more cEEGs with reactivity (good outcome: 4 of 6 vs. bad outcome: 1 of 12, p = 0.02) and sleep features (good outcome: 5 of 6 vs. bad outcome: 2 of 12, p = 0.01) following anakinra initiation.

Conclusions: The absence of HEBs following anakinra treatment may signal anakinra response; thus, a potential marker to start anesthetic withdrawal. cEEG background features may provide early insight into the eventual neurological recovery.

Funding: None