Withdrawing antiseizure drugs before five seizure-free years was linked to more relapse

Design

Systematic review and meta-analysis of observational studies

Population

22 studies, 3,933 patients withdrawing antiseizure medication

Primary outcome

Seizure relapse after withdrawal by seizure-free period

Effect

Under 5 years seizure-free: RR 1.49 (95% CI 1.17 to 1.88)

A meta-analysis in the Journal of Neurology, published on 20 October 2026, identified 45 studies of antiseizure medication withdrawal and pooled 22 of them, covering 3,933 patients, on the question of how long the seizure-free period should be before stopping.

Patients withdrawn after fewer than five seizure-free years had a higher risk of relapse than those withdrawn later (RR 1.49, 95% CI 1.17 to 1.88). In subgroups, children with under three years and adults with under five years of seizure freedom also had higher relapse risk. The authors note that estimates for very long seizure-free periods rest on few patients.

The findings are observational and mix different populations and withdrawal protocols. They support the shared decision rather than fix a threshold: a longer seizure-free period lowers the odds of relapse, but the right moment still depends on epilepsy type, EEG, imaging and what a seizure would cost the patient.

• Discuss the length of seizure freedom as one input when a patient asks about stopping medication.
• Use a validated individual prediction tool alongside seizure-free duration where available.
• Discuss driving and work consequences of a relapse before withdrawal.
• Taper gradually and agree a plan for what happens if a seizure recurs.

Why it matters

Patients often ask to stop at two years; this suggests the risk at that point is higher than many assume.

Don't overread it

Pooled observational studies with varied protocols — it does not set a universal waiting period.

The statistics, in plain English

A relative risk of 1.49 means relapse was about half as likely again with a shorter seizure-free period. Subgroup estimates by age and region are less reliable than the overall result.

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.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.