Blood thinners cut strokes and clots by 69% in a trial of 1,803 heart patients

In a randomized trial of 1,803 patients with atrial fibrillation, direct oral anticoagulants cut the trial’s primary composite endpoint by 69% over 24 months, from 1.5% of patients with no anticoagulation to 0.5% of those on a blood thinner. The composite counts stroke, systemic embolism, major bleeding and cardiovascular death together, and ischemic stroke drove most of the gap: 0.1% on drug against 1.1% without it.

The trial, called SINGLE-AF, enrolled patients at 18 medical centers in South Korea and was led by Professor Boyoung Joung of Yonsei University in Seoul, with Daehoon Kim of the same institution among the principal authors. It was presented in a Hot Line session at ESC Congress 2026 and published in the New England Journal of Medicine.

A gap in the evidence for intermediate-risk patients

Patients with atrial fibrillation and a high stroke-risk score are routinely anticoagulated; those with a low score are not. The middle band was the problem. SINGLE-AF enrolled people with a CHA2DS2-VASc score of 1 in men or 2 in women, and the European Society of Cardiology’s press release notes that earlier guidance gave anticoagulation only a Class IIa, meaning “should be considered,” recommendation for these patients because randomized data were lacking.

Participants were randomized 1:1 to apixaban 5 mg twice daily or rivaroxaban 20 mg once daily, or to no anticoagulation. According to TCTMD’s coverage, mean age was 60.4 years and 71.8% had paroxysmal atrial fibrillation. The design registered as NCT04437654 in the published protocol was open-label with blinded outcome assessment, and the primary endpoint was changed in January 2025 from all-cause death to cardiovascular death after the data monitoring board observed that most deaths up to then were non-cardiovascular.

Reading the 69% as a composite

The 69% is a relative reduction in a four-part composite, with a hazard ratio of 0.31 and a 95% confidence interval of 0.10 to 0.94 (p=0.028), as reported in the repost of the ESC release and by TCTMD. The wide interval is the visible sign of how few events there were: the lower bound implies a 90% reduction and the upper bound a 6% one. Ischemic stroke, the largest single driver, occurred in 0.1% of patients taking DOACs and 1.1% of those taking nothing.

The distinction between the composite and its parts is the one a reader needs to hold onto. The 69% does not describe strokes and clots alone; it describes the first of any four events, including two that anticoagulants can cause or that have nothing to do with clotting. Stroke is the component that moved: an absolute difference of 1.0 percentage point over two years, 0.1% against 1.1%. Systemic embolism, the other clot-related event, sits inside the same composite, and the published summaries do not break it out separately. The clean message is that fewer patients on a DOAC reached any of the four endpoints, mainly because fewer of them had an ischemic stroke.

Because major bleeding sits inside the composite, the safety result matters as much as the efficacy one. Major bleeding occurred in 0.3% of the DOAC group and 0.5% of the control group, so the anticoagulated patients did not bleed more in this trial.

Event counts were low throughout, which is the central limitation. With absolute rates of 0.5% and 1.5% across two years, a difference of a handful of patients shifts the percentages, and the trial’s own authors acknowledge it. As TCTMD reports, they wrote that the difference “was based on a small number of events and may overestimate the true magnitude of benefit,” and that confirmatory studies are needed.

Independent caution and the population tested

Rod Passman, an independent expert quoted by TCTMD, said: “This is not a game changer.” He pointed to the small number of events, the possibility that the effect is specific to the Korean population studied, and the lifestyle burden of lifelong anticoagulation for patients whose absolute risk is low. Those reservations sit alongside Joung’s own framing of the result; the ESC release quotes him saying SINGLE-AF data “may be used to inform future guideline recommendations and reimbursement policies.”

The trial tested two specific drugs in one health system over 24 months. Whether the same margin holds in other populations or ethnic groups, at longer follow-up, or against a strategy such as rhythm control or catheter ablation is outside what 1,803 patients can establish. The primary paper in the New England Journal of Medicine is where those details are tabulated, and its 95% interval, running from a 90% reduction down to a 6% one, is the number that guideline writers will have to weigh.

This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.

More from Morning Overview