Background: Cervical artery dissection is a major cause of stroke in young people (aged <50 years). Historically, clinicians have preferred using oral anticoagulation with vitamin K antagonists for patients with cervical artery dissection, although some current guidelines-based on available evidence from mostly observational studies-suggest using aspirin. If proven to be non-inferior to vitamin K antagonists, aspirin might be preferable, due to its ease of use and lower cost. We aimed to test the non-inferiority of aspirin to vitamin K antagonists in patients with cervical artery dissection.

Methods: We did a multicentre, randomised, open-label, non-inferiority trial in ten stroke centres across Switzerland, Germany, and Denmark. We randomly assigned (1:1) patients aged older than 18 years who had symptomatic, MRI-verified, cervical artery dissection within 2 weeks before enrolment, to receive either aspirin 300 mg once daily or a vitamin K antagonist (phenprocoumon, acenocoumarol, or warfarin; target international normalised ratio [INR] 2·0-3·0) for 90 days. Randomisation was computer-generated using an interactive web response system, with stratification according to participating site. Independent imaging core laboratory adjudicators were masked to treatment allocation, but investigators, patients, and clinical event adjudicators were aware of treatment allocation. The primary endpoint was a composite of clinical outcomes (stroke, major haemorrhage, or death) and MRI outcomes (new ischaemic or haemorrhagic brain lesions) in the per-protocol population, assessed at 14 days (clinical and MRI outcomes) and 90 days (clinical outcomes only) after commencing treatment. Non-inferiority of aspirin would be shown if the upper limit of the two-sided 95% CI of the absolute risk difference between groups was less than 12% (non-inferiority margin). This trial is registered with ClinicalTrials.gov, NCT02046460.

Findings: Between Sept 11, 2013, and Dec 21, 2018, we enrolled 194 patients; 100 (52%) were assigned to the aspirin group and 94 (48%) were assigned to the vitamin K antagonist group. The per-protocol population included 173 patients; 91 (53%) in the aspirin group and 82 (47%) in the vitamin K antagonist group. The primary endpoint occurred in 21 (23%) of 91 patients in the aspirin group and in 12 (15%) of 82 patients in the vitamin K antagonist group (absolute difference 8% [95% CI -4 to 21], non-inferiority p=0·55). Thus, non-inferiority of aspirin was not shown. Seven patients (8%) in the aspirin group and none in the vitamin K antagonist group had ischaemic strokes. One patient (1%) in the vitamin K antagonist group and none in the aspirin group had major extracranial haemorrhage. There were no deaths. Subclinical MRI outcomes were recorded in 14 patients (15%) in the aspirin group and in 11 patients (13%) in the vitamin K antagonist group. There were 19 adverse events in the aspirin group, and 26 in the vitamin K antagonist group.

Interpretation: Our findings did not show that aspirin was non-inferior to vitamin K antagonists in the treatment of cervical artery dissection.

Funding: Swiss National Science Foundation, Swiss Heart Foundation, Stroke Funds Basel, University Hospital Basel, University of Basel, Academic Society Basel.