- Design
- Systematic review and meta-analysis of randomised trials
- Population
- 8 trials, 1,000 children with new-onset or inadequately controlled epilepsy
- Primary outcome
- Seizure freedom
- Effect
- RR 0.95 (95% CI 0.83 to 1.09); I² 59.5%; low certainty
A meta-analysis in the European Journal of Clinical Pharmacology, published on 20 October 2026, pooled eight randomised trials of 1,000 children with new-onset epilepsy or epilepsy inadequately controlled on earlier treatment, comparing levetiracetam with valproate.
Seizure freedom did not differ significantly (five trials, 655 children; RR 0.95, 95% CI 0.83 to 1.09), but heterogeneity was moderate and the result depended on one trial, so certainty was low. Exploratory time-to-recurrence analyses favoured valproate, though censoring was poorly handled. Levetiracetam was associated with fewer adverse events overall and less weight gain; behavioural effects and liver safety remained uncertain.
The practical reading is that neither drug clearly beats the other on seizure control in these trials, so the choice rests on the child. For girls approaching child-bearing age, the known teratogenic and neurodevelopmental risks of valproate in pregnancy remain the dominant consideration, independent of this analysis.
- Choose between levetiracetam and valproate by epilepsy type, age, sex and side-effect profile rather than expected seizure control.
- Avoid valproate in girls who may become pregnant unless no alternative works, and follow pregnancy prevention requirements.
- Ask about mood and behaviour changes at follow-up on levetiracetam.
- Monitor weight on valproate.
- For generalised epilepsies, check whether syndrome-specific guidance favours valproate in boys.
Why it matters
It supports levetiracetam as a reasonable first choice where valproate's long-term risks weigh heavily.
Don't overread it
Low-certainty evidence that rests on one trial; it does not establish equivalence across epilepsy syndromes.
The statistics, in plain English
A relative risk of 0.95 with an interval of 0.83 to 1.09 is compatible with either drug being slightly better. I-squared of about 60% means the trials disagreed more than chance would explain, which lowers confidence in the pooled figure.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for paediatrics, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free