A rigorous sham-controlled trial has failed to demonstrate a significant benefit for patent foramen ovale (PFO) closure in addressing frequent migraine headaches.
There has long been a strong suspicion that some migraine headaches may be caused by PFO, but this relationship has never been conclusively demonstrated, and it has never been established that PFO closure can reduce migraine attacks.
The PREMIUM (Prospective, Randomized Investigation to Evaluate Incidence of Headache Reduction in Subjects With Migraine and PFO Using the AMPLATZER PFO Occluder to Medical Management) trial studied patients with frequent migraines who had failed medical therapy and had a significant right-to-left shunt on transcranial Doppler.
Two hundred and thirty patients were randomized to a sham procedure or PFO closure with the St. Jude Medical Amplatzer PFO occluder device.
The trial did not show a significant reduction in the primary efficacy endpoint (a 50% reduction in migraine attacks), which was reached in 38.5% of patients in the device arm and 32% of patients in the control arm.
The device was generally safe, although one patient in the PFO closure group had transient atrial fibrillation, as reported in the Journal of the American College of Cardiology.
However, there were several indications that PFO closure may ultimately prove to be helpful in some patients. 8.5% of patients in the device arms had a complete cessation of migraine attacks, compared to only 1% in the control arm.
There was a significant signal of benefit for PFO closure in the subset of patients with consistent or frequent migraine aura.
“PFO closure is not a cure for migraine to be applied broadly, but may be an important therapy for some,” wrote Brian Whisenant (Intermountain Medical Center, Salt Lake City) and Mark Reisman (University of Washington), in an accompanying editorial.
“Given the tremendous unmet need of additional migraine prevention therapies, the safety of PFO closure, and ongoing observations of migraine improvement in some patients, future research must focus on removing the blinders and identifying those who may be most responsive to PFO closure.”
In an extensive interview, the lead author of the study, Jonathan Tobis (UCLA) discussed the long and troubled history of PFO closure for migraine. He noted that the results of PREMIUM were first presented 2 years ago at the TCT meeting but that it took several years to get the paper published because of “a lot of resistance in the neurology community.”
The neurology community was also “very resistant to the idea that cryptogenic stroke was related to embolism,” he said.
Tobis believes that PFO closure will eventually be found to be beneficial in some patients, but does not think that migraine patients should now receive PFO closure outside the setting of a randomized trial.
“Ultimately, we will be able to demonstrate in randomized controlled trials that a certain subset of migraines is also causally related to the right-to-left shunt PFO,” he predicted.
Tobis said that he estimates that about half of patients with migraine with aura have a PFO. The goal, he said, is to identify a migraine that is due to the PFO and that will respond to PFO closure.
Tobis said that talks are underway for a new trial that may use a novel method to identify which patients are more likely to respond to PFO closure. “I am optimistic that there will be another randomized trial with a better defined subset of people who are more likely to respond.”
Tobis also talked about the importance of using sham controls. “The placebo effect is much larger than anyone predicted.”
Tobis also addressed the problem of recruiting patients in interventional cardiology trials in general and PFO closure trials in particular: “Although there is presumed equipoise for a device versus medical therapy, most patients who are willing to participate want the device.
This was addressed in PREMIUM by allowing the patients to receive the PFO closure device after the blind was broken at 1 year. This was at no cost to the patient and thus is an inducement to be randomized … 80% of control subjects asked for and received the device. They were followed for 1 year for any safety concerns, but the effect on their migraine was not studied since they were no longer blinded at that point.”
Tobis said he believes that the FDA should allow patients who participate in device trials to receive the device when the trial is over, even if the device is still not approved at that time.
He also said that companies should “budget for free device procedures in the control group after the blind is broken. This will encourage subject participation and also is ethically fair to the control population.”