After appendage occlusion, low-dose rivaroxaban beat antiplatelet therapy on brain lesions

Design

investigator-initiated, prospective, multicentre, randomised controlled trial with serial diffusion-weighted MRI (HALO-SCE)

Population

164 patients with successful left atrial appendage occlusion confirmed at 45 days, all eligible for oral anticoagulation

Primary outcome

patient-level incidence of any newly detected silent cerebral embolic lesion during follow-up to 365 days

Effect

12.2% (10/82) with rivaroxaban 10 mg vs 31.7% (26/82) with antiplatelet therapy, P=0.005; MMSE difference 2.56 (95% CI 1.11 to 4.01) at 365 days

What to give after a successful left atrial appendage occlusion has never been settled. HALO-SCE randomised 164 patients, 45 days after occlusion was confirmed successful, to rivaroxaban 10 mg daily or to antiplatelet therapy, with diffusion-weighted MRI and cognitive testing repeated at 90, 180 and 365 days. All were eligible for oral anticoagulation.

New silent cerebral embolic lesions occurred in 10 of 82 (12.2%) on half-dose rivaroxaban and 26 of 82 (31.7%) on antiplatelet therapy (P=0.005). Cognitive scores at 365 days favoured rivaroxaban on both instruments used: a between-group difference of 2.56 points on the Mini-Mental State Examination (95% CI 1.11 to 4.01) and 2.67 on the Montreal Cognitive Assessment (1.07 to 4.26). The composite of death, clinical thromboembolism and major bleeding was 2.4% against 11.0%, which did not reach significance (P=0.057).

The device is implanted precisely so that anticoagulation can stop, so a trial showing that continuing a reduced dose does better is worth attention. But read the endpoints for what they are: silent lesions on MRI are a surrogate, and while the cognitive difference is the more clinically meaningful finding, a 2.5-point shift on these scales in an open imaging-driven trial of 164 patients is not the same as demonstrated dementia prevention. The clinical composite pointed the same way and did not reach significance. This argues for the question being taken seriously in a larger trial, and for reconsidering the reflex that occlusion means antiplatelet therapy thereafter in a patient who could take an anticoagulant.

• Note the population: patients still eligible for oral anticoagulation, not those in whom bleeding forced the occlusion.
• The primary endpoint is imaging - new silent embolic lesions - not stroke.
• Cognitive differences of about 2.5 points are the clinically interesting signal, and rest on 164 patients over one year.
• The clinical composite favoured rivaroxaban at 2.4% versus 11.0% but did not reach significance at P=0.057.
• Do not generalise to patients occluded because anticoagulation was contraindicated - they were not studied here.

The statistics, in plain English

Silent cerebral embolic lesions are a surrogate: they are associated with cognitive decline but are not themselves an event a patient experiences. The cognitive results are more meaningful precisely because they are patient-level, though a difference of 2.56 MMSE points with an interval of 1.11 to 4.01 in an unblinded-to-treatment population needs replication before it is quoted as fact. The composite clinical outcome, 2.4% against 11.0% with P=0.057, is a good illustration of why a trial of 164 patients cannot settle clinical events - that is a large relative difference sitting just outside significance because the absolute numbers are tiny.

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.