A multicentre Chinese trial of drugs for migraine with a PFO
Migraine affects roughly one billion people worldwide and is among the most common disabling neurological conditions. About one in four adults has a patent foramen ovale (PFO) — a small channel in the wall between the heart’s upper chambers that does not fully close after birth — and PFO is more common in people with migraine than in the general population. One explanation, the “micro-embolism hypothesis,” is that the channel lets tiny clots slip past the lungs into brain arteries and trigger attacks.
Fuwai Hospital, the national centre for cardiovascular diseases, reports that a team led by its principal investigator Professor Pan Xiangbin ran the COMPETE study (Comparison of the Effect of Medication in Alleviating Migraine with Patent Foramen Ovale) across 39 Chinese hospitals. The team enrolled 1,000 patients with PFO-related migraine and randomised them to one of four oral treatments: aspirin, clopidogrel, the direct oral anticoagulant rivaroxaban, or the beta-blocker metoprolol, which is a common migraine preventive. The BMJ published the results (BMJ 2026;394:e100103) and commissioned a companion editorial by the international headache and cerebrovascular-disease expert Gretchen E. Tietjen (BMJ 2026;394:e100359).
What the trial measured and found
The primary outcome was the proportion of patients whose migraine days or attacks fell by at least 50% from baseline at 12 weeks.
Aspirin 61.7%, clopidogrel 66.8%, rivaroxaban 78.4%, metoprolol 61.8%. All three antiplatelet or anticoagulant options were at least non-inferior to metoprolol; rivaroxaban was significantly more effective than metoprolol overall, and the advantage was clearest in patients with aura and in those with a large right-to-left shunt (grade 3). The investigators interpret this as consistent with a micro-embolism mechanism and as a possible functional marker for who may benefit later from PFO closure.
What is still unknown
The 12-week primary endpoint measured symptom response, not long-term migraine control, stroke, or bleeding events over years. Patients were not blinded to treatment, the comparator was a single preventive drug, and the trial did not include a placebo arm. The companion editorial notes that, although the design was pragmatic, the absence of placebo means the absolute size of the benefit should be read with caution. How these drugs compare with newer migraine-specific preventives (such as CGRP antagonists) in PFO-positive patients is not addressed here. The study also does not test whether a positive drug response actually predicts who will later benefit from closing the PFO — that is the planned next step.
What this may mean for patients
For adults with migraine and a confirmed PFO who are not already taking an anticoagulant, the trial offers randomised evidence that a low-dose direct oral anticoagulant is a reasonable preventive option, with a higher chance of halving migraine days at 12 weeks than a standard beta-blocker. This is not a recommendation for any individual to start, stop, or switch a medicine; the original paper and a clinician familiar with the patient’s full history are the right places to weigh the choice, including bleeding risk. The trial is also part of a planned sequence: the Fuwai team has registered two follow-up studies, COMPETE-Cohort and COMPETE-BioMetal, to test PFO closure against continued medical therapy and to evaluate a fully bioabsorbable closure device.
Relevance to migraine and stroke-related cardiac care
Migraine with aura sits in Wellward China’s Brain & Nervous System health area, and PFO sits in Heart & Circulation; this trial sits across both, because it is testing a heart-side intervention for what is experienced as a brain condition. The same underlying PFO is also a known route for paradoxical embolism that can cause cryptogenic stroke, which is why cardiology, neurology and stroke medicine are all involved.