PFO Closure for Migraine: Pooled Analysis Hints at Benefit Over Meds Alone

Patent foramen ovale (PFO) closure in conjunction with optimal medical therapy can reduce both the number of migraine days and the frequency of attacks compared with medicine alone, according to a new pooled analysis.

The findings are unlikely to have an effect on clinical practice, as regulatory bodies have not approved any PFO closure device for the treatment of migraine. However, the procedure is now an option for the prevention of PFO-associated stroke.

“What this [study] will do is increase people's interest and whet their appetite for the next clinical trial, but it's not going to have an immediate practical, clinical effect,” senior study author Jonathan Tobis, MD (University of California, Los Angeles), told TCTMD. “People aren't going to go out and start closing PFOs for migraine. I hope if anything it will stimulate interest in the patient population to participate in the RELIEF PFO trial [because] this shows the good justification for why PFO closure may be beneficial in a certain subset.”

Recruitment for the randomized RELIEF trial—the first of its kind to use thienopyridines as a screening mechanism to identify migraine patients who would be more likely to benefit from PFO closure with the Cardioform Septal Occluder (Gore)—has been delayed by the COVID-19 pandemic, but it will be starting this year.

A Hole New Option?

For their study, published in the February 16, 2021, issue of the Journal of the American College of Cardiology, Tobis along with Mohammad Mojadidi, MD (Virginia Commonwealth University, Richmond), Preetham Kumar, MD (University of California, Los Angeles), and colleagues combined individual patient-level data from both the PRIMA and PREMIUM trials comparing medical therapy with and without closure with the Amplatzer PFO Occluder (Abbott) for migraine relief. Both randomized trials failed to meet their primary endpoints—total migraine days and responder rate (defined as ≥ 50% decrease in migraine attacks with or without aura), respectively.

Taken together, however, the findings from 337 patients show benefit for PFO closure over medical therapy alone with regard to mean reduction in monthly migraine days (-3.1 vs -1.9; P = 0.02), mean reduction of monthly migraine attacks (-2.0 vs -1.4; P = 0.01), and number of individuals who reported complete cessation of their migraines (9% vs 0.7%; P < 0.001). There was no advantage seen for PFO closure in terms of responder rate (38% vs 29%; P = 0.13).

Looking only at patients who have migraine with aura, PFO closure significantly reduced the number of migraine days (-3.2 vs -1.8; P = 0.03) and was associated with more reports of total headache cessation (11% vs 0.9%; P = 0.002). By contrast, neither of these endpoints were significantly different among patients with migraine without aura.

In total, PFO closure led to nine procedure-related events, including access-site hematoma and transient hypotension, and four device-related adverse events, with the most common being paroxysmal atrial fibrillation.

Tobis said researchers in this space have been frustrated. “All of us in this field have seen many patients who say that their migraines get better when the PFO is closed, and there are lots of observational studies that document this,” he observed. “But those are usually for people with less severe migraines, less frequent migraines. And then the randomized trials for severe migraine did not show clinical benefit, yet PRIMA and PREMIUM came pretty close.”

However, he lamented, if the PREMIUM trial was able to use reduction in total migraine days as its primary endpoint, it would have been a positive trial.

With regard to the neutral results for responder rate, Tobis said he views it as more of a “complicated” statistical question. “If you have 10-20 migraine days per month, it's easier to show a reduction in half than if you have four migraine attacks and you have to show a reduction from four to two,” he said. “It's harder to show a statistical significance. So that's why I believe that the responder rate is less likely to demonstrate a beneficial response and is not as good a demonstration or predictor of clinical trial results.”

Reinvigorating the Question

“This patient level meta-analysis is worthwhile as it reinvigorates the question of whether PFO closure could be a viable therapeutic option to reduce the burden of migraines in some patients,” Steven Messé, MD (Perelman School of Medicine at the University of Pennsylvania, Philadelphia), who was not involved in the study, told TCTMD in an email. “I don't think the results are that surprising as the combined results are fairly concordant with the results of the two studies, but with increased precision and power and a focus on the more encouraging secondary outcomes.”

However, he disagrees somewhat with Tobis, stressing that even if PREMIUM had been allowed to use reduction in total migraine days as its primary endpoint, there would not have been much of a difference in clinical implications. “The clinical impact of a difference in the reduction of monthly migraine attacks of only 0.6, and only 1.4 fewer migraine days is of uncertain clinical importance,” he said, noting that responder rate is typically seen as a more meaningful metric. “Importantly, this meta-analysis did not show a significant increase in patients achieving this benchmark, although it did favor closure numerically and likely was underpowered to show a difference. I think it is also worth pointing out that all of the outcomes evaluated in this meta-analysis were selected in retrospect.”

“Clinically, I do not think we should be doing anything at this point for patients with migraine and a PFO, other than encouraging them to participate in randomized trials if they are eligible so that we can learn whether closure is effective and for which patients,” Messé said. While he is “not aware” of any migraine patients getting PFO closure off-label, “I would not be surprised if it is happening,” he added.

Pathological Mechanism Missing

In an editorial accompanying the study, Zubair Ahmed, MD (Cleveland Clinic, OH), and RELIEF primary investigator Robert Sommer, MD (Columbia University Irving Medical Center, New York, NY), write that while the researchers did show benefit with PFO closure for migraine in this pooled analysis, “perhaps the most fundamental limitation of this and all PFO-migraine trials to date is that the underlying pathophysiologic mechanism linking migraine symptoms to PFO remains unknown.”

Many have speculated that right-to-left shunt might play a role, as it does in PFO-related stroke, though this has yet to be proven. “But as experience and trials have shown, some migraine patients receive no benefit from PFO closure,” they write. “As the majority of migraineurs have no PFOs, they must have ‘PFO-unrelated’ migraines and must have a pathophysiologic mechanism or trigger different from that associated with the right-to-left shunt.”

Further research must better identify patients with “causal” versus “incidental” PFOs and not simply focus on patients with and without aura, according to Ahmed and Sommer. “Designing a randomized trial to prove the PFO migraine mechanism beyond a reasonable doubt will require an inclusion or screening criterion that can reliably determine the causal or incidental nature of the PFO. Only then will the benefit of PFO closure therapy be truly knowable (the inclusion of only causal PFO would likely also significantly increase the rate of beneficial response to the therapy).”

To be included in the RELIEF trial, patients must show a 50% reduction in monthly migraine days after taking P2Y12 inhibitors. From there, they will undergo a blinded PFO closure or sham procedure, and the thienopyridine approach will be validated with a separate randomized, blinded placebo arm during the initial phase of the trial.

“Once we learn to identify the subset of migraine patients in whom PFOs are actually causal of headache symptoms, screening and treatment of PFO for migraine can become a reality,” the editorialists conclude.

Messé, to TCTMD, applauded the trial design. “It is encouraging that they are trying to select the patients most likely to benefit from closure in the RELIEF trial, and performing sham procedures for the controls. Clearly, there is a large placebo effect in migraine trials,” Messé said. “In the meantime, since these PFO closure trials were completed, a new class of migraine medicines—the CGRP inhibitors—have been approved and are being widely used, hopefully reducing the burden for patients who suffer with frequent migraines.”