Abstract
Background:
The association of patient foramen ovale (PFO) and migraine headache (migraine) with aura (MA) is well established. Current research suggests a mechanistic link between platelet activation, paradoxical embolization and migraine in some patients.
Methods:
Clopidogrel, a platelet inhibitor, was added to existing migraine therapy, as a 4-week open-label trial in 15 women, aged 16–56 years, with severe migraine and documented right to left shunt (RLS).
Results:
13/15 had >50% reduction or complete elimination of migraine symptoms. After completing the trial period, five responders remain on clopidogrel with ongoing benefit at 11.9 ± 4.5 months (6.5–20), one stopped clopidogrel because of side effects. Nine other responders underwent PFO closure and clopidogrel discontinuation. Eight of nine have had ongoing benefit.
Conclusions:
Clopidogrel may have a primary prophylactic role in migraine/RLS patients, but may also help select candidates who would benefit from PFO closure. A randomized clinical trial is being established.
Introduction
The association of patent foramen ovale (PFO) and migraine headaches with aura (MA) has been well established. It has been demonstrated that 40–50% of patients with MA have PFO, significantly different from the migraine population without preceding aura (M0), and the non-migraine population, which are both 20–25% (1). However, it is far from established, nor universally accepted, that there is a direct causal relationship, as not all patients with PFO have migraine, and not all migraine patients have a PFO. Current research suggests a link between migraine, platelet activation, paradoxical embolization, and hypercoagulable states (2).
In our interventional cardiology practice, a subset of PFO/stroke patients with aspirin sensitivity or allergy had been treated with clopidogrel, a thienopyridine platelet inhibitor, as primary stroke preventative therapy pending, or in lieu of PFO closure. In some, co-existing migraine symptoms improved while taking the drug, prior to PFO closure. This prompted us to consider testing clopidogrel as a novel prophylactic therapy for severe migraineurs, without stroke history, who also had a documented right to left shunt (RLS). We report the results in our first 15 patients.
Patients/methods
Approval was obtained from the Columbia University IRB for a chart review of patients treated, in our Cardiology Practice, between June 2010 and June 2013. Patients with chronic, severe migraine symptoms, and known RLS, who had been treated with clopidogrel therapy, were included. The referring neurologist had confirmed the diagnosis of both migraine and RLS prior to our consultation. Preceding aura was not used as a selection criterion. There was no predetermined maximum number of monthly migraine days. A minimum of two headache-days/week was required to allow us to detect a change within a short window of therapy. Patients with a history of transient ischemic attack or stroke were excluded.
Patient and headache characteristics, prior therapy, and effect of trial drug.
CM: chronic migraine; M0: migraine headache without aura; RLS: right to left shunt; PFO: patent foramen ovale; MA: migraine with aura; TCD: transcranial Doppler; TEE: transesophageal echocardiogram; TTE: transthoracic echocardiogram.
Results
Patients ranged from 16 to 56 years of age (mean 32.3 ± 11.9 years). All were female. All had chronic, severe symptoms, many having failed multiple prior therapies (see Table 1). All had documented RLS by transcranial Doppler (12), by transesophageal echocardiogram (2), or by transthoracic echocardiogram (1) studies. Ten patients met criteria for episodic migraine, five for chronic migraine (CM). Seven of the 15 patients had visual scotoma, six had atypical aura symptoms, two had both visual scotoma and atypical aura features, and one had basilar symptoms. Three of the patients were M0.
Thirteen of the 15 patients experienced immediate improvement on clopidogrel (‘responders’) with at least a 50% reduction in headache frequency from their baseline. Nine of the 13 responders had no migraine symptoms while on the therapy. One (Patient 4, MA) had no benefit. One (Patient 13, MA) seemed to have an initial resolution of symptoms during the trial month and for 1 month after PFO closure, and then had return of migraine symptoms while still on clopidogrel. There were insufficient data to assess a change in headache severity, in patients who had ongoing, but less frequent symptoms.
Five of the responders remain on clopidogrel with ongoing benefit at a mean follow-up of 11.9 ± 4.5 months (range 6.5–20.0 months). One has tried stopping the drug several times, each time with return of symptoms. One needed to discontinue the medication because of apparent allergic symptoms.
The other nine initial responders elected to undergo PFO closure following the 4-week trial. Patient 13 had return of migraine within 1 month (as above). Patient 8 had resumption of migraines after discontinuation of clopidogrel 3 months after closure, but again had complete elimination of symptoms on restarting the medication. The clopidogrel was again discontinued 3 months later. The patient remains headache-free since. Patient 6, who had had daily symptoms at baseline, had no migraines after PFO closure except with extreme exercise. Resumption of the clopidogrel, at 6 months, eliminated the remaining symptoms. The remaining six closure patients have had ongoing improvement or elimination of migraine symptoms, after discontinuation of clopidogrel.
Discussion
Wilmshurst (3) first reported unanticipated improvement in migraine symptoms after transcatheter device closure of an atrial communication. A number of subsequent worldwide observational studies reported improvement or elimination of migraine in patients who had had a PFO closed for prevention of recurrent embolic stroke (4). At that time it was proposed that the RLS, in some way triggered the headaches. This correlated well to increased migraine prevalence in patients with cyanotic congenital heart disease (5) and with hereditary hemorrhagic telangiectasia (6), both of whom also have RLS. In the latter report, a reduction in migraine frequency was noted after transcatheter elimination of the RLS.
The MIST Trial (7) was the first double blinded, randomized, sham-controlled study of PFO closure in 147 patients with severe episodic MA and a documented PFO with moderate or large RLS. M0 patients, those with chronic migraine, and those with a history of stroke, were excluded. This controversial study ultimately showed no benefit to PFO closure, casting further uncertainty on the causative role of the PFO.
Subsequent work has focused on potential underlying pathophysiologic mechanisms linking migraine and PFO. Using cerebral microembolization in a mouse model, Nozari et al. (2), were able to produce cortical-spreading depression, a finding linked with MA in humans. Their work clearly demonstrated the potential of cerebral paradoxical embolization to trigger migraine without stroke.
Similarly, in humans, migraine have been triggered both by venous injection of agitated saline (8) during transcranial Doppler and echocardiographic studies, and by foam sclerotherapy for venous insufficiency (9).
Wilmshurst et al. (10), noted an increased frequency of MA symptoms in the first few weeks after PFO closure. This effect was mitigated by the use of clopidogrel, suggesting a causative role of platelet activation, in those patients. The timing of headache resolution, in this series and in others, correlated to the endothelialization of the PFO closure device, which may take 3–6 months. This may explain the transient return of migraine symptoms in our Patient 8.
An alternative pathophysiologic mechanism may link migraine headaches and venous nitric oxide levels (11). Platelet activation has been closely linked to modulation of the endothelium via nitric oxide and other mediators. Endothelin, for example, a potent vasoconstrictor, is present in high levels in the venous circulation after injections of sclerotherapy agents (10) and is normally broken down on passage through the pulmonary circulation. With a RLS, vasoactive substances could bypass the lungs, reaching the cerebral circulation in high concentrations, and trigger migraine symptoms. Clopidogrel, by inhibiting the release of these agents, would potentially reduce cerebral concentrations. Animal models have shown clopidogrel to modulate vasoconstriction (12) in vivo.
Prior migraine/PFO trials have focused exclusively on the patient population with episodic migraine. In the MIST Trial (7), the ESCAPE Migraine Trial (St. Jude Medical, never completed/published), and the PREMIUM Trial (active/no longer enrolling – St. Jude Medical/AGA Medical), chronic daily migraines were excluded. If paradoxical passage of micro-emboli or of vasoactive substances is the mechanism for migraine in some, it is unclear why these patients would not have daily headaches, as the PFO opens continuously. Perhaps there is a minimum embolic ‘load’ required to trigger the event. However, our Patients 5, 6, and 8, with chronic migraine symptoms, also seemed to respond well to clopidogrel therapy.
There are obvious limitations to this report. In an open-label/unblinded series, we cannot rule out a placebo effect of the therapy. Patients were specifically told that it might improve their headaches. However, the profound benefits of the therapy, and its duration in all but one, were remarkable and inconsistent with a placebo effect.
The month-to-month variability of episodic migraines could also have misled us to attribute benefit to a therapy that may have been incidental. A longer study period is required to ensure that the results are truly causal.
The results of any retrospective study are always limited by the inherent inaccuracy of the patient's recall of symptoms, although, in this case, a dramatic reduction in headache symptoms would be easily recalled over the course of a short 4-week trial period.
Conclusions
There may exist a subset of migraine patients in whom RLS is causal for migraine. For patients with severe migraine who have failed, or cannot use conventional migraine therapies, screening for RLS, and in those found to have a RLS, therapy with a thienopyridine agent, may prove beneficial to some. As suggested by these preliminary data, clopidogrel responsiveness might also be a predictor of migraineurs who would benefit from PFO closure.
Based on these preliminary observations, we are establishing a prospective, multi-center, randomized, double-blinded, placebo-controlled trial to assess the effectiveness of thienopyridine therapy (The THEOREM Trial – THiEnOpyridine for Refractory Episodic Migraine), as the true benefit of any such treatment can only be defined in a formal randomized study.
Clinical implications
Clopidogrel may be a novel prophylactic migraine therapy for patients with right to left shunt. Clopidogrel responsiveness may potentially predict migraineurs whose triggers are related to a right to left shunt. Clopidogrel responsiveness may potentially predict migraineurs who would benefit from PFO closure.
Footnotes
Acknowledgement
The authors wish to thank the South Shore Neurology group for their ongoing support and encouragement in this area of investigation.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Conflict of interest
Dr. Sommer reports the following conflicts of interest: 1) Stock options in Coherex Medical, a company that makes a PFO closure device; 2) Local Primary Investigator at Columbia University Medical Center for the REDUCE Trial, sponsored by GORE Medical; and 3) Previously a National Primary Investigator for the ESCAPE Migraine Trial sponsored by St. Jude Medical. The other authors have no conflicts of interest to disclose.