Antiseizure drug choice linked to thrombosis and bleeding in patients on DOACs

Design

Retrospective target-trial emulation with propensity matching, federated electronic records

Population

9529 adults with epilepsy starting antiseizure monotherapy while on a DOAC

Primary outcome

Thromboembolic composite, major bleeding and all-cause mortality

Effect

Versus lamotrigine/lacosamide: strong inducers thromboembolism HR 1.55 (95% CI 1.02–2.36); levetiracetam HR 1.98 (1.37–2.87); valproate intracranial bleeding HR 3.01 (1.45–6.26)

This study emulated a target trial using electronic records from 165 healthcare organisations. It followed 9529 adults with epilepsy who started antiseizure monotherapy while taking a direct oral anticoagulant, and compared each drug group with lamotrigine or lacosamide after propensity matching.

Strong enzyme inducers such as carbamazepine and phenytoin were associated with more thromboembolic events (HR 1.55) and less major bleeding. That fits their known effect of lowering DOAC levels. Valproate was associated with higher mortality and three times the rate of intracranial bleeding. Levetiracetam, which has no important DOAC interaction, was also associated with more thromboembolism (HR 1.98) and higher mortality. That is harder to explain biologically and may reflect who receives levetiracetam, such as frailer patients after a stroke.

In patients with epilepsy who need anticoagulation, avoid strong enzyme inducers. This matters in India, where phenytoin and carbamazepine remain widely used. Lamotrigine and lacosamide are reasonable first choices. The levetiracetam signal is worth noting but should not drive a switch on its own.

• Avoid carbamazepine and phenytoin in patients who need a DOAC
• Consider lamotrigine or lacosamide when starting an antiseizure drug alongside a DOAC
• Be cautious with valproate where intracranial bleeding risk is high
• Review the whole drug list for other CYP3A4 and P-glycoprotein interactions

Why it matters

The choice of antiseizure drug may be a modifiable contributor to stroke and bleeding in anticoagulated patients.

Don't overread it

This was observational — the levetiracetam association in particular may reflect which patients are prescribed it, not a drug effect.

The statistics, in plain English

HR 1.55 (95% CI 1.02–2.36) for enzyme inducers just excludes no effect; HR 1.98 (1.37–2.87) for levetiracetam is stronger. Propensity matching balances measured factors only, so unmeasured differences, such as frailty or reasons for choosing a drug, can still produce these associations.

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.