Levetiracetam Versus Valproate in Children: What the Evidence Shows

What was studied

Researchers combined results from 8 randomized controlled trials involving 1,000 children ages 0–18 years. The children had either new-onset epilepsy or epilepsy that had been inadequately controlled by previous treatment.

The review compared levetiracetam with valproate. Outcomes included seizure freedom, survival time without a seizure, time to the first seizure recurrence, AST levels, weight gain, and other adverse events.

What they found

Seizure freedom did not differ significantly between levetiracetam and valproate. Across 5 studies with 655 participants, the relative risk was 0.95 (95% CI 0.83–1.09; P=0.45; I²=59.5%). This was rated as low-certainty evidence.

Exploratory analyses of survival time without a seizure and time to the first seizure recurrence favored valproate, but these findings were also low certainty and were limited by inadequate handling of censoring.

Levetiracetam was associated with fewer overall adverse events and less weight gain. It was also associated with lower AST levels, but this did not establish that levetiracetam was safer for the liver. Behavioral and hepatic safety remained uncertain.

Limits of the evidence

The seizure-freedom evidence was low certainty, and the finding was sensitive to one trial. The reported I² of 59.5% indicated inconsistency among study results. The time-related analyses were exploratory and did not adequately handle censoring. The abstract does not provide treatment durations, doses, epilepsy types, the sizes of the adverse-event differences, or detailed behavioral outcomes.

For families and caregivers

In these trials, neither medicine clearly resulted in better seizure freedom overall. Levetiracetam was associated with fewer overall adverse events and less weight gain, while exploratory seizure-timing results favored valproate. Because the evidence is uncertain, treatment should be individualized according to patient- and medication-related factors rather than assuming one medicine is best for every child.

What to watch next

Further research should confirm the seizure findings, use appropriate methods for censoring in time-related analyses, and clarify behavioral and hepatic safety.

Terms in this summary

Randomized controlled trial

A study in which participants are assigned by chance to different treatments to make the comparison fairer.

Seizure freedom

Having no seizures during the study’s specified follow-up period.

Relative risk (RR)

A comparison of how often an outcome occurs in two groups; an RR near 1 suggests little or no difference.

95% confidence interval (CI)

A range showing uncertainty around an estimate. Here, the range included 1, so a difference in seizure freedom was not established.

I²

A measure of inconsistency among study results; the reported value was 59.5%.

AST

Aspartate aminotransferase, an enzyme measured as part of the study’s safety assessment.

Censoring

A time-to-event issue in which the exact time of an outcome is not known for every participant.

Low-certainty evidence

Evidence for which confidence in the estimated effect is limited.