By Jeffrey Zimmet, MD, PhD

Associate Professor of Medicine, University of California, San Francisco; Director, Cardiac Catheterization Laboratory, San Francisco VA Medical Center

Dr. Zimmet reports no financial relationships relevant to this field of study.

SYNOPSIS: This randomized trial of patent foramen ovale closure in severe refractory migraine showed no significant difference in responder rate compared to sham control.

SOURCE: Tobis JM, Charles A, Silberstein SD, et al. Percutaneous closure of patent foramen ovale in patients with migraine: The PREMIUM Trial. J Am Coll Cardiol 2017;70:2766-2774.

Migraines are common. Patent foramen ovale (PFO) is common. Numerous observational studies have suggested an association between PFO and migraine, especially for migraine with aura. Similarly, observational data previously have supported the idea that closure of right-to-left shunt, including PFO, can reduce the frequency and severity of migraines. However, not every person with migraines has a PFO, and not every PFO is associated with migraines.

The key to a successful trial of PFO closure in migraines is to identify the subset of migraine patients who are most likely to benefit. This key point partially explains the success of PFO closure in three recent trials that included patients with cryptogenic stroke: RESPECT, Gore REDUCE, and CLOSE, in which the enrollment of younger patients with significant shunts in whom other causes of stroke had been excluded resulted in measurement of a significant treatment effect.

The PREMIUM trial was envisioned as the migraine counterpart to these trials. The authors of PREMIUM enrolled highly symptomatic patients with between six and 14 headache days per month who had failed at least three different preventive medications. All patients were required to have a large shunt, assessed as grade 4 or 5 (more than 100 bubbles per minute) by transcranial Doppler. Although prominent aura had been identified as a possible predictor of a positive treatment effect of PFO closure, limiting enrollment to this group turned out to be impractical; thus, patients both with and without aura were included. Despite this, the authors needed more than seven years to enroll the 230 patients required. Out of 1,653 patients who were screened and consented, 1,423 did not meet the inclusion criteria; most of these did not have a large enough shunt. In the end, 123 patients were randomized to receive a PFO closure device, while 107 patients were randomized to the control group. All enrolled patients underwent right heart catheterization while under deep sedation. Randomization was performed only after the PFO was crossed with a wire. Those randomized to intervention received an Amplatzer PFO Occluder, which was implanted successfully in all but four patients. Both the subjects and their treating neurologists were blinded to the treatment assignment.

After 12 months of follow-up, there was no significant difference between the device arm and the sham controls in the primary efficacy endpoint: 38.5% of device patients and 32% of the control subjects demonstrated a 50% reduction in migraine attacks (P = 0.32). Among secondary endpoints, patients in the device arm showed a statistically significant decrease in mean migraine days per month, but the reported difference was very modest: -3.4 ± 4.4 days vs. -2.0 ± 5.0 days (P = 0.025). Interestingly, 10 of 117 patients in the device arm stopped experiencing migraine attacks completely vs. only one control patient. Six of the 10 complete responders were patients with aura. In a subgroup analysis that included only the patients with aura (79 patients out of the trial population of 230), the responder rate was 49% in device patients vs. 23% for controls (P = 0.015).

The authors reported that PFO closure in patients with migraine and significant shunt did not change the responder rate significantly. Therefore, their results do not support the use of PFO closure as a preventive therapy for migraine. The authors noted that the significant results in the subgroup of patients with aura leaves open the possibility of a role for this therapy in a subset of patients.


Migraines can be incredibly debilitating, especially when the symptoms are frequent and not responsive to available therapies. The PREMIUM trial, like other studies before it, gives tantalizing evidence that there is the possibility of a treatment effect for PFO in migraines, if only we could identify the patients who would benefit. In the overall population of the trial, there was a measurable and significant difference in migraine days per month, but the magnitude of this effect was quite small.

More encouraging were the results in the subset of patients with migraine with aura, who make up a minority of all patients with migraine. This trial was relatively enriched in this population. An optimist would say that there is a link between migraine with aura and PFO, and the requirement of a large shunt selected more of these patients into the study. However, the fact that this was a non-prespecified subgroup means that these findings are at best hypothesis-generating and may form the basis for future studies.

In the same vein, the finding that more patients in the device arm achieved complete remission of migraine over a year of follow-up is encouraging and generates hope of successful therapy for some proportion of migraine patients. However, the percentage of patients who achieved this result (8.5%) was too small to suggest PFO closure as a non-investigational treatment today. Undoubtedly, some patients will continue to be offered PFO closure for this indication, but until better data indicate which patients are likely to benefit, this will remain on the fringes of clinical practice.