Half-dose rivaroxaban after appendage occlusion cut silent cerebral emboli compared with antiplatelets

Design

Investigator-initiated, multicentre, randomised controlled trial

Population

164 patients after successful left atrial appendage occlusion, eligible for oral anticoagulation

Primary outcome

New silent cerebral embolic lesions on diffusion-weighted MRI over 365 days

Effect

12.2% vs 31.7% (P = 0.005); MoCA +2.67 (95% CI 1.07–4.26) at 1 year

HALO-SCE randomised 164 patients with a successful left atrial appendage occlusion (confirmed at 45 days) to rivaroxaban 10 mg daily or antiplatelet therapy. Diffusion-weighted MRI and cognitive tests were repeated at 90, 180 and 365 days.

New silent cerebral embolic lesions occurred in 12.2% (10 of 82) on half-dose rivaroxaban and 31.7% (26 of 82) on antiplatelet therapy (P = 0.005). At one year, cognitive scores favoured rivaroxaban (MMSE difference 2.56, 95% CI 1.11 to 4.01; MoCA 2.67, 1.07 to 4.26). A composite of death, clinical thromboembolism and major bleeding was 2.4% vs 11.0% (P = 0.057).

The trial is small, single-country and built on an imaging surrogate. It enrolled patients who were eligible for anticoagulation, which is not the usual reason for choosing occlusion. It does not answer what to do in patients who had the device because they cannot take anticoagulants.

• After occlusion in a patient who can tolerate anticoagulation, low-dose rivaroxaban is an off-label option supported so far only by imaging and cognitive endpoints; follow the device protocol and current guidance.
• The evidence rests on MRI lesions and cognitive scores, not on strokes or deaths.
• Do not extrapolate to patients who had occlusion because of prior major bleeding.
• Confirm the device seal on follow-up imaging before choosing the long-term regimen.

Why it matters

It questions whether antiplatelet therapy alone is enough protection after appendage occlusion.

Don't overread it

Silent lesions and test scores are surrogates; the trial was not powered for strokes, bleeding or death.

The statistics, in plain English

With 82 patients per arm, a difference of 10 vs 26 patients with new lesions is statistically clear, but the clinical composite (2.4% vs 11.0%) has P = 0.057, just short of the usual threshold. Small trials can overstate effects.

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.