Drug resistance and clinical course in children and youth with epilepsy: a population-based cohort
Truls Vikin, Morten I Lossius, Ragnhild E Brandlistuen, Dag Hofoss, Richard F Chin, Kari M AabergAbstract
In this large, population-based cohort with validated childhood-onset epilepsy, we aimed to provide contemporary estimates of drug-resistant epilepsy and to delineate predictors and seizure-course patterns using the latest International League Against Epilepsy classifications.
Children and youth from the Norwegian Mother, Father and Child Cohort Study (base population approximately 114 500)—with clinically validated epilepsy and ≥2 years of follow-up were included (sample size 1020; 53% male; onset 0–18 years; median onset age 6.0 years (interquartile range 2.3–9.8); median age at last follow-up 16.7 years (interquartile range 14.9–18.4)). Drug-resistant epilepsy was defined as persistent seizures in the final follow-up year despite trials of two or more antiseizure medications. Predictors were assessed with log-binomial models (crude) and Poisson regression with robust errors (multivariable), adjusting for follow-up time and investigation indicators; syndrome-specific analyses and five seizure-course categories (favourable/improving/poor/deteriorating/varying) complemented the binary endpoint.
At last follow-up, 19% (194 of 1020) met criteria for drug-resistant epilepsy, consistent with long-term population estimates. Independent associations included three or more seizure types, identified or presumed genetic aetiology, intellectual disability, status epilepticus, younger than two years at onset, reported seizure precipitants, and non-epileptiform electroencephalographic abnormalities. Adolescent-onset was not significant after adjustment. Across syndromes, developmental and epileptic encephalopathies (including Lennox–Gastaut syndrome, Dravet syndrome and infantile epileptic spasms syndrome) carried the highest risk, whereas self-limited epilepsies and childhood absence epilepsy showed excellent long-term outcomes. Seizure-course patterns reflected these risks: for infantile epileptic spasms syndrome, seizure control was rare and typically achieved early if at all, whereas self-limited epilepsy with centrotemporal spikes was almost uniformly non-drug-resistant yet often followed a delayed, ultimately favourable course.
These contemporary, population-based findings show that one in five children with epilepsy meets the criteria for drug resistance despite modern diagnostics and treatment. Risk is strongly influenced by the number of seizure types, aetiology, developmental profile, status epilepticus and electroencephalographic background. The observed association with reported precipitants merits replication and may aid risk stratification alongside syndrome and aetiology.