Following the disappointing results of the MIST Trial, physicians in Europe explored the hypothesis that use of a more effective PFO closure device might improve the clinical results in patients with migraine and aura. The PRIMA Trial used the Amplatzer PFO occluder device, which is reported to have a higher complete closure rate and a lower complication profile than the StarFlex device. The study screened 2000 patients who had migraine with aura. Using agitated saline contrast, trans-thoracic echocardiography revealed that 40% of their patient population had evidence for a right-to-left shunt through the atrial septum. Unlike the MIST Trial, PRIMA was not a blinded or sham controlled trial. Unfortunately, since patients were randomized before the date of the procedure, a number of patients dropped out of the trial after randomization, so that 12 out of 53 subjects (23%) assigned to the closure group, did not get a device. In addition, it was very difficult to recruit patients due to the strict criteria of being unresponsive to 2 preventive medications for migraine. The study was therefore stopped early by the sponsor, St. Jude Medical.

The primary endpoint of the trial was the change in the total number of headache days per month, both with and without aura. In the 107 patients with a mean age of 44 years who completed the trial, the primary endpoint was not met. In the device closure group consisting of 53 people, the mean number of headache days at baseline was 8.0 and decreased to 5.1 +/– 4.7 days/month. In the medical treated arm, there were 54 patients who had 8.3 headache days at baseline and this decreased to 6.6 +/– 2.4 days/month, p = 0.17. However, the secondary endpoints were achieved with the number of headache with aura days significantly reduced (p = 0.01) as well as the number of attacks (P < 0.01). PFO closure was also associated with a greater responder rate (≥50 % reduction in number of migraine days) compared with medical therapy (37.5 % vs. 14.6 %; P = 0.02). More closure patients reported freedom from migraine (10 % vs. 0%) and freedom from migraine with aura (40 % vs. 10 %) than those treated with medical therapy (P < 0.05 for both).

In conclusion, although the PRIMA Trial did not achieve its primary endpoint, there was evidence for a significant improvement in headaches for the subjects who had frequent migraine with aura. Due to the inability to randomize an adequate number of subjects, the PRIMA Trial was underpowered and suffered from a high dropout of people. The lack of blinding and a sham procedure were also detrimental aspects of this trial design. These issues were addressed in the PREMIUM Trial, which used the same PFO occluder device.

The PREMIUM trial (St. Jude Medical, St. Paul, Minnesota) is a double-blind sham-controlled study to evaluate the effect of PFO closure in patients with migraine headaches using the Amplatzer PFO occluder. In contrast to the MIST trial, the PREMIUM study includes subjects with and without aura, and without regard to the age at which the first onset of migraine headaches occurred. Another aim of the inclusion criteria was to exclude people with chronic daily headaches (more than 15 headache days per month) since these subjects often take excessive medication and may have rebound headaches from withdrawal of the pain medications. Tension headaches, visual aura without headache, or probable migraines were counted as headache days. Based on review of a 60-day screening migraine diary, subjects with 5–14 headache days per month, who historically had failed 3 or more classes of preventive medications, were enrolled in the trial.

To screen patients for the PREMIUM Trial, right-to-left shunts were quantified with a TCD bubble study. A subject had to have a TCD grade 4 or 5 (more than 100 bubbles in 1 min) to be enrolled. Subjects who met the inclusion criteria were then brought into the catheterization lab and were heavily sedated, but not placed under general anesthesia (as used in the MIST Trial). They wore headphones to prevent hearing any operator conversations and their visual pathway to the video monitors was blocked. The operator first had to prove that there was a PFO by passing a guidewire across the atrial septum under visualization by intracardiac echo. This avoided the discomfort of a screening TEE for making the diagnosis of a PFO and also eliminated any subject where the PFO could not be crossed. The aim was to prevent operators from performing trans-septal punctures to place the device so that the risk of the closure procedure could be reduced. This feature was proposed in response to the MIST Trial which had a number of complications associated with the procedure itself. The subjects kept a diary to record the frequency of headaches for 3 months before the procedure, and for 1 year after randomization. At the end of 1 year, the number of headache days was counted from the daily diary for the last 3 months of observation and was compared to the baseline number of headache days. Evidence of a residual right-to-left shunt was documented by TCD and TTE. Only subjects with a TCD grade greater than 3 after closure received a TEE to confirm that the residual shunt was at the level of the atrial septum and not within the pulmonary vasculature.

Subjects were unblinded at the end of the first year. Those people who were assigned to medical therapy alone were given an option to have their PFO closed. This unusual design was requested of the Food and drug administration to encourage subjects to participate in a randomized trial that had a sham procedure arm during the treatment phase, even though the device had not been approved yet. Data on patients in the optional arm was used to obtain more information about device safety; however, since the subjects were unblinded, any information about the effect of PFO closure on their migraines was not used to demonstrate efficacy.

Currently, the PREMIUM trial has closed subject enrollment and approaches the data analysis phase. It will take 1 year after the last subject was enrolled to obtain the headache frequency data. There are concerns that by including subjects who have migraine without aura into the study population, the efficacious results may get diluted as the observational studies demonstrate that headaches in patients without aura do not improve as frequently as in those who have migraine with aura. However, the observational studies did show a moderate response even in people who had migraine without aura once their PFO was closed, so the decision was made not to exclude these patients from the PREMIUM trial. In one study, 31 % of migraineurs without aura reported complete relief of their migraines once the PFO was closed, and an additional 38 % of patients said their migraines were reduced by half [5].

As of this writing date, the PREMIUM trial is currently in progress. Although all 230 patients have been enrolled and randomized, the data collection is to continue until June 2014, when the headache frequency data from the last subject enrolled in the study becomes available. It remains to be seen whether PFO closure with the Amplatzer PFO device will demonstrate a different result from the MIST Trial where the STARFlex device was used. The data will also be stratified so we will also learn if migraineurs without aura respond differently than migraineurs with aura. It is possible that the study population in both the MIST and PREMIUM Trials are less responsive than the patients with cryptogenic stroke who also had migraine. For those physicians who are working in this field and for the millions of people who suffer from severe migraine headaches, the results of the PREMIUM Trial are described in the Sect. The PRIMA Trial.