A rare form of stroke leaves patients with months of blood thinners and a choice of tablet. A registry of nearly 2,500 people adds data to that choice and shows how far evidence without randomization can take it.
On 19 October 2026 the journal Stroke published an analysis of 2,476 adults from a multicentre Chinese registry of cerebral venous thrombosis, a blood clot in the veins that drain the brain. All had the clot confirmed on imaging and started an oral anticoagulant within 30 days of diagnosis: 942 a direct oral anticoagulant (DOAC) — rivaroxaban in 701, dabigatran in 121, apixaban in 113 and edoxaban in 7 — and 1,534 warfarin.
Within a year, a new symptomatic clot in a vein anywhere in the body occurred in 36 people on a DOAC and 73 on warfarin. After statistical weighting to make the groups comparable, the one-year risk was 3.74% with a DOAC and 5.47% with warfarin, a difference of −1.73 percentage points (95% CI −3.76 to 0.30). The authors had set in advance that a DOAC would count as "not worse" if its risk was no more than 1.5 percentage points higher than warfarin's; that criterion was met (P = 0.004). The lower risk itself was not statistically significant: the subdistribution hazard ratio was 0.67 (95% CI 0.37–1.21; P = 0.184).
Major bleeding was rare: 0 patients on a DOAC and 1 on warfarin. Clinically relevant non-major bleeding occurred in 18 and 34 patients, and any bleeding in 48 and 75. The authors state that the bleeding comparisons were exploratory and do not allow conclusions about which drug is safer.
A trial built from registry data
The study is a target trial emulation: the researchers wrote down the protocol of the randomized trial they would have wanted and then reproduced it with observational data. Follow-up started on the day of the first anticoagulant prescription, the groups were balanced by propensity-score overlap weighting, and deaths were treated as a competing event, so that a person who died was not counted as someone who simply stayed clot-free. Patients came from a retrospective phase in 2015–2022 and a prospective phase in 2023–2025.
A rare clot with little randomized evidence
Vitamin K antagonists such as warfarin have traditionally been used after cerebral venous thrombosis, while DOACs are prescribed more and more often. Randomized comparisons so far have been small. RE-SPECT CVT, published in JAMA Neurology in 2019, included 120 patients followed for 25 weeks: no recurrent venous clots occurred in either group, and there was one major bleed with dabigatran and two with warfarin. The authors of that trial called it exploratory, and a Canadian randomized feasibility trial of rivaroxaban, published in Stroke in 2023, enrolled 55 participants. With such small numbers, a registry of almost 2,500 people adds weight to the question patients face after the acute phase: which tablet to take for the following months.
Where the evidence stops
Treatment was not assigned at random. Doctors chose the drug, and weighting can only balance the factors that were recorded; differences that were not measured may remain. The analysis shows that DOACs were not worse than warfarin by the chosen margin, not that they were better. Three quarters of the DOAC group took one drug, rivaroxaban, so the result says little about apixaban and almost nothing about edoxaban. The data come from Chinese centres, and the abstract does not report funding or conflicts of interest. The authors give no regulatory timeline; a larger randomized trial would be the next step for a firm answer.
If you or a relative take an anticoagulant after a cerebral venous thrombosis, the choice of drug and how long to take it are questions to discuss with your doctor; switching on your own is not safe.
This material is for information only and does not replace a consultation with a doctor. Nothing in it is a recommendation to start, stop or change treatment.
Sources
Zhong J, et al. Initial Direct Oral Anticoagulants Versus Warfarin After Cerebral Venous Thrombosis: A Target Trial Emulation. Stroke. 2026. PMID 42895027
Field TS, et al. Study of Rivaroxaban for Cerebral Venous Thrombosis: A Randomized Controlled Feasibility Trial. Stroke. 2023. PMID 37675613
Ferro JM, et al. Safety and Efficacy of Dabigatran Etexilate vs Dose-Adjusted Warfarin in Patients With Cerebral Venous Thrombosis: A Randomized Clinical Trial. JAMA Neurol. 2019. PMID 31479105
What this piece relies on
Cards
Articles in this section are educational and are not medical advice, a diagnosis, or a prescription. Consult a qualified professional before acting on anything you read here.
Read next
ResearchMedicine and treatmentAfter a heart-valve implant, the frailest patients benefited most
Frailty usually marks the patients doctors worry about most after a heart procedure. A new analysis of a randomized trial suggests it may also point to who gains the most from treatment, which matters to older people facing a valve implant and to their families, with clear limits on what it shows.
4 min read
ResearchMedicine and treatmentLeukaemia without chemotherapy at the start: what a phase 3 trial found
Blood cancer, type 2 diabetes, small thyroid cancers and ageing research: four recent results, each with the population or model it was measured in and the point where the evidence stops.
6 min read
ResearchMedicine and treatmentUS approves a tablet for herpes lesions unresponsive to antivirals
This issue examines difficult-to-treat herpes in people with weakened immunity, evidence behind decisions in bladder cancer, links between brown fat and cardiometabolic health, and early research on personalised ovarian cancer vaccines. It separates clinical results from extrapolation and association.
6 min read