In a retrospective cohort study published in Acta Epileptologica, researchers followed 100 people whose epilepsy began before the age of 12 and who had been under care for at least ten years. By the end of the observation window, only 12 percent had achieved seizure freedom, and statistical testing found no significant improvement in seizure control between the start of treatment and the ten-year mark.
The study was carried out at the adult and paediatric epilepsy clinics of Hospital Canselor Tuanku Muhriz, Universiti Kebangsaan Malaysia, between June 2021 and December 2022. The team screened 117 patients and included 100 after excluding 17 whose records lacked key data. Participants were recruited through universal sampling, and diagnoses followed the criteria of the International League Against Epilepsy. The researchers deliberately excluded people whose seizures were febrile or provoked by acute causes such as encephalitis, toxins, severe electrolyte imbalances, or space-occupying lesions like brain tumours or abscesses. Data came from manual extraction of medical records, supplemented by history taking during follow-up visits.
The cohort was strikingly young at onset. The mean age at epilepsy diagnosis was 3.97 years, with the youngest patient diagnosed at ten months and the oldest at twelve years. By the time of the study, participants ranged from 11 to 46 years old, with a mean age of 23.15 years and an even split between sexes. Focal seizures dominated the clinical picture, accounting for 53 percent of cases, followed by generalised epilepsy at 31 percent. Structural causes were the leading aetiology, while 33 percent of cases had an unknown cause and 13 percent were genetic. The most common imaging findings were malformations of cortical development, seen in 30 percent of patients, and hypoxic-ischemic encephalopathy, seen in 11 percent.
The burden of the disease extended far beyond seizures themselves. Nearly half of the participants, 48 percent, had an intellectual disability, 24 percent had a learning disability, and only 28 percent were intellectually normal. Behavioural difficulties affected 32 percent, with autism spectrum disorder at 10 percent and attention deficit hyperactivity disorder at 6 percent among the most common diagnoses. More than half of the cohort, 58 percent, were formally registered as disabled.
Electroencephalography findings painted a consistent picture of drug-resistant disease. Before treatment, interictal EEG showed focal epileptiform discharges in 68 percent of patients, generalised discharges in 12 percent, and normal results in just 11 percent. After antiseizure medication was started, more than half of patients, 56 percent, showed no abnormalities on EEG. Focal discharges fell from 68 cases at baseline to 35 after treatment, a normalisation rate of 48.5 percent, while generalised discharges dropped from 12 cases to 9, a normalisation rate of 25 percent. Ictal EEG recordings were unavailable for more than half of the patients, but among those who had them, focal epileptiform discharges predominated at 28 percent.
Treatment patterns reflected long-standing clinical practice. Over one-third of participants, 36 percent, received dual antiseizure medication therapy, 34 percent were on polytherapy, and 29 percent were on monotherapy. Carbamazepine and sodium valproate were each prescribed to 49 percent of patients, making them by far the most common agents. The median number of medications per patient was two, with a range of one to seven. Adherence was high, with 82 percent of patients compliant with treatment, and 76 percent reported no adverse effects; among those who did, drowsiness, dizziness, and weight gain were each reported by 3 percent. A small minority, 8 percent, received non-pharmacological treatments, including ketogenic diet therapy in 4 percent, epilepsy surgery in 3 percent, and neuromodulation in 1 percent.
The statistical analysis told a sobering story about the trajectory of seizure control. A Wilcoxon signed-rank test comparing seizure control from baseline through ten years of follow-up found no statistically significant difference, with a Z statistic of −0.577 and a P value of 0.564. At baseline, seizure freedom was significantly associated with seizure aetiology, behavioural issues, evidence of perinatal asphyxia, and the type of antiseizure medication used. After ten years, however, antiseizure medication type was the only variable that retained a significant relationship with seizure freedom. The authors then ran a logistic regression to identify predictors of seizure freedom, but the overall model failed to reach statistical significance, with a chi-square of 29.299 on 19 degrees of freedom, a P value of 0.061, and a Nagelkerke R-squared of 0.531. None of the individual predictors, spanning aetiology, behaviour, and medication, were significant.
The authors are candid about why their remission figures fall well below the two-thirds terminal remission rates reported in much of the international literature. Because patients with self-limited childhood epilepsies tend to stop attending long-term follow-up once their seizures resolve, the final cohort was skewed towards drug-resistant cases. The single-centre design and retrospective method further limit how far the findings can be generalised to the wider community. Of the 117 patients screened, 17 were excluded for missing key data, leaving the study slightly below its calculated target of 105 participants, which the authors acknowledge may have limited the statistical power of some subgroup analyses. Reliance on documented medical records also meant the team could not confirm pre-existing cognitive impairment or changes in seizure semiology over time, introducing a degree of selection bias. The authors call for prospective, multicentre research involving other local and regional hospitals, with follow-up extended to at least 20 years.
Despite these constraints, the clinical picture that emerges is consistent: seizure semiology remained stable over the decade, focal epilepsy predominated among cases persisting into adulthood, and structural causes accounted for roughly half of all aetiologies. The researchers link this pattern to the enduring impact of early structural insults such as hypoxic-ischemic injuries and focal cortical malformations, which they argue explains why chronic epilepsy persisting beyond childhood is disproportionately focal rather than self-limiting. Their cohort also had a lower proportion of unknown aetiologies than most comparable studies, which they attribute to the advanced imaging and genetic testing available at their tertiary centre, though cost remains a barrier to pursuing these investigations more widely.
From these findings, the team proposes a structured, three-phase transition pathway for young patients moving from paediatric to adult neurology services. In the preparation phase, from ages 12 to 15, allied health assessments including clinical psychology screening for psychiatric issues would take place, alongside mandatory neuroimaging to exclude surgically treatable lesions, particularly in drug-resistant cases. In the joint assessment phase, from ages 16 to 18, paediatric and adult neurologists would hold joint consultations to build trust, with an epilepsy nurse coordinating care and education and introductions to local non-governmental organisations. In the final adult assimilation phase, after age 18, formal transfer to adult neurology would secure continuity of long-term care. The authors argue that this multidisciplinary model, incorporating clinical psychologists, epilepsy nurses, and social or advocacy support, is essential given the heavy burden of intellectual, learning, and behavioural disabilities their cohort carried, and that early screening for anxiety and depression could allow prompt intervention with meaningful clinical impact.
Subject of Research: Long-term clinical outcomes of childhood-onset epilepsy followed into adulthood at a Malaysian tertiary centre
Article Title: A decade later: a longitudinal retrospective cohort study of the clinical course of childhood-onset epilepsy in a multi-ethnic tertiary centre
Article References: Tan, J. E., Tan, J. K., Tan, M. M., Ng, Z. X., Chan, Z. Y., Ang, Z. Q., Mohd Khaili, N. A., Rahmat, N. L., Lee, J., Khoo, C. S., Tan, H. J., & Wong, S. W. (2026). A decade later: a longitudinal retrospective cohort study of the clinical course of childhood-onset epilepsy in a multi-ethnic tertiary centre. Acta Epileptologica, 8(1), Article 34. https://doi.org/10.1186/s42494-026-00277-z
Image Credits: AI Generated
DOI: 10.1186/s42494-026-00277-z
Keywords: epilepsy, childhood-onset epilepsy, seizure control, retrospective cohort study, neurodevelopmental disability, antiseizure medication, EEG, drug-resistant epilepsy, paediatric-to-adult transition, tertiary care, Malaysia, long-term outcomes
News Source: Kendall Mcintyre. (October 28, 2026). Childhood epilepsy often persists into adulthood with little improvement over a decade. Scienmag.
