Papers of interest – October 2026

What they did. Researchers reviewed the records of 1,213 children seen with new seizure-like events at a tertiary first seizure clinic in the Netherlands. Every child had a routine EEG at the first visit. The team then looked at which later tests helped reach a final diagnosis.

What they found. At the first visit, 23% of children were diagnosed with epilepsy, 43% were found not to have epilepsy and 34% remained undecided. By follow-up, 34% had epilepsy, 61% did not and 6% were still unclear; only 1.9% of children had been misclassified at the first visit. Among the 934 children without an initial epilepsy diagnosis, a repeat EEG contributed to a final diagnosis of epilepsy in 7.7%, MRI in 1.2% and genetic testing in 0.4%. In children whose diagnosis was initially unclear but who turned out to have epilepsy, a sleep-deprived EEG showed epileptiform discharges in up to 71%.

Bear in mind. This was a retrospective study from one specialist clinic, and clinicians chose which children had which tests, so the figures describe yield in practice rather than the accuracy of each test.

What they did. A children’s hospital in China reviewed 628 children with focal cortical dysplasia seen on MRI between 2006 and 2025, looking at who developed epilepsy, who became drug-resistant and how children did after surgery.

What they found. 79% of the children developed epilepsy and 24% developed drug-resistant epilepsy. Of 81 children who had surgery and were followed up for the medium to long term, 73% were completely seizure-free. Developmental delay, intellectual disability, motor impairment, underdeveloped white matter and some interictal EEG patterns were linked to developing epilepsy. The children fell into three broad groups, from low risk (normal intellect, a lesion in a single lobe and a normal EEG) to high risk (drug-resistant epilepsy, intellectual disability and lesions across several lobes).

Bear in mind. This was a retrospective study from a single centre. The children who had surgery were selected and there was no comparison group, so the seizure-freedom figure is not a treatment effect or an individual prognosis.

What they did. As part of a US project to identify Lennox–Gastaut syndrome automatically from electronic health records, seven paediatric epilepsy centres reviewed 935 children coded with Lennox–Gastaut syndrome or other intractable epilepsy between 2016 and 2022. Two epileptologists independently classified each child as classic, pragmatic or not Lennox–Gastaut syndrome, with a senior epileptologist settling disagreements.

What they found. The two reviewers disagreed in 11% of cases. 130 records had to be re-reviewed to find missing EEG reports and clinical details, and this changed the classification of 67 children first judged not to have the syndrome: 33 were reclassified as classic and 12 as pragmatic Lennox–Gastaut syndrome. The commonest problems were inconsistent EEG reporting, incomplete descriptions of seizure types, scattered information about cognition, unclear age at onset and varying terminology. Slow spike-and-wave findings were the element most often missing.

Bear in mind. The study was designed to build research cohorts from US health records, so its lessons are about documentation rather than how often the diagnosis is right in clinic.

What they did. A single centre reviewed 90 children who had their first epilepsy surgery before the age of 6. Development was measured on a standard developmental scale, and the rate of progress in the first year after surgery was compared with the rate afterwards.

What they found. Children developed faster in the first year after surgery than in later years (median developmental rate 0.22 compared with 0.08). This pattern held regardless of the epilepsy syndrome, type of operation, cause, development before surgery, how long the child had had epilepsy, or whether seizures came back in the first year. Faster early progress was linked to better development before surgery and an improved EEG at one year. The authors suggest the early boost may reflect relief from the effects of frequent seizures, while later progress depends more on the underlying cause.

Bear in mind. This was a retrospective study from one centre using a single developmental measure, with no comparison group of children who did not have surgery. The authors describe their findings as hypothesis-generating.

What they did. Researchers at the University of Leicester interviewed 17 people in 10 interviews: seven children aged 5 to 12 with their carers, and three carers on their own. Families were recruited through epilepsy charities and interviewed online.

What they found. Four themes emerged: how children and parents think and feel about the medicines; organising and managing them day to day; the formulation and how it is given; and support from family, school and professionals. Children mainly raised the formulation, especially the taste. Parents focused more on side effects and the importance of routines. Some parents of children with a learning disability were also concerned about the colour of the medicine. The authors recommend that professionals ask about formulation problems early and explain clearly what to expect from treatment.

Bear in mind. This was a small qualitative study of families recruited through charities, so it may not reflect the experience of every family.

What they did. This short report used US national death records from 1999 to 2023 for children and young people aged up to 19 whose underlying cause of death was epilepsy or status epilepticus, and looked at how death rates changed over time by age, sex and ethnicity.

What they found. There were 4,472 deaths over the period. The age-adjusted death rate rose from 0.2 to 0.4 per 100,000, increasing by about 5.6% a year from 2001 (95% CI 3.1% to 13.7%). Among 11 to 19 year olds, rates have risen by about 8.6% a year since 2008. Rates were stable overall in children aged 10 and under, but rose in children aged 2 and under. The authors suggest SUDEP in adolescence and a growing burden of other conditions as possible explanations, and note that SUDEP is often under-recognised.

Bear in mind. The study relies on US death certificate coding, which can change over time, and could only analyse Black and White non-Hispanic groups. It shows trends but cannot explain them, and the figures may not apply to the UK.