Percutaneous closure of patent foramen ovale reduces the frequency of migraine attacks

Methods.

In March 2003, all patients with presumed PFO-mediated paradoxical embolism referred to our center for percutaneous PFO closure between January 1998 and March 2002 were asked to participate. The local ethics committee approved the study, and patients gave written informed consent.

The clinical diagnosis of presumptive paradoxical embolism was based on the absence of a left-sided thromboembolic source and the presence of a PFO-mediated right-to-left shunt as seen on transesophageal contrast echocardiography. To rule out a left-sided source of thromboembolism, color-coded Doppler and duplex examinations of the carotid and vertebral arteries, transcranial Doppler of the intracranial arteries, and EKG were performed in all patients with cerebral ischemic events. Cerebral ischemia was verified by CT or MR and in some patients also by cerebral angiography.

Percutaneous PFO closure was performed according to the device-specific instructions for use. After closure, all patients received acetylsalicylic acid 100 mg once daily until transesophageal echocardiography at 6 months revealed no or only a minimal residual shunt. Clopidogrel 75 mg/day was added to aspirin for 1 month after the intervention.

Patients answered a detailed questionnaire addressing medical history and recurrent embolic events during follow-up. Headache frequency and characteristics were assessed retrospectively for the year before and after percutaneous PFO closure.

Headache was classified by a neurologist as migraine with aura, migraine without aura, tension-type headache, and other or not classifiable headache according to the International Headache Society criteria. Tension-type headache and other or not classifiable headache were subsumed as nonmigraine headache in our study.

Predictors for the improvement of headache symptoms after PFO closure were calculated using a multivariate logistic regression analysis with stepwise backward elimination.

Results.

Between January 1998 and March 2002, 275 patients underwent percutaneous PFO closure, 202 patients (73%) within the last 2 years. Two hundred fifteen patients (78%) agreed to participate in the study. Clinical characteristics of patients with (n = 71) and without (n = 144) headache the year prior to percutaneous PFO closure are summarized in table 1. Indications for percutaneous PFO closure were cerebral (n = 175; 81%) and peripheral (n = 13; 6%) embolism, decompression illness (n = 26; 12%), and aggravated hypoxemia in pulmonary embolism (n = 1). Indications did not differ between patients with and without headache. Six months after percutaneous PFO closure, 167 patients (78%) underwent repeat transesophageal contrast echocardiography at our institution (see table 1).

The overall 1-year prevalence of headache prior to percutaneous PFO closure was 33% (n = 71). Nonmigraine headache was diagnosed in 23 patients (11%), and 48 patients (22%) fulfilled the criteria for migraine (table 2). Visual symptoms (98%) were the most frequent aura manifestation, followed by sensory (22%), aphasic (14%), and motor symptoms (3%). There were no differences between patients with migraine and nonmigraine headache with respect to duration of headache history (25 ± 14 vs 20 ± 13 years; p = NS) and frequency of headache attacks (1.2 vs 1.4 per month; p = NS).

Eight patients (11%) were not able to classify the effect of percutaneous PFO closure on headache symptoms. After percutaneous PFO closure, the frequency of headache attacks was reduced by 54% (p = 0.001) in patients with migraine with aura and by 62% (p = 0.006; see table 2) in patients with migraine without aura. Patients with nonmigraine headaches did not experience a decrease in the frequency of headache attacks after the intervention (see table 2). A residual shunt after the intervention was present in 3 of 7 patients (43%) with migraine with aura who failed to improve but only in 1 of 12 patients (8%) with nonmigraine headache who did not improve (p = NS).

Logistic regression analysis revealed migraine with and without aura as the only predictor for improvement of headache symptoms after percutaneous PFO closure (odds ratio [OR] 5.2, 95% CI 1.3 to 20). Female gender (OR 0.9, 95% CI 0.3 to 3.0) and duration of headache history (OR 0.99, 95% CI 0.94 to 1.03) were not predictive.

Discussion.

The current study adds further evidence for an association between PFO and migraine. First, in patients with presumed PFO-mediated paradoxical embolism, the prevalence of migraine with aura was twice the expected prevalence in the general population. Second, percutaneous PFO closure reduced the frequency of migraine attacks with and without aura but not that of nonmigraine headache. Third, the diagnosis of migraine headache was the only predictor for improvement of headache symptoms after percutaneous PFO closure.

Migraine is most widely perceived as an increased individual excitability of neuronal depolarization, probably determined by genetic and environmental factors.4 In the laboratory setting, the aura phenomenon preceding migraine headache can be initiated by hypoxia, hypoglycemia, mechanical trauma, and changes in the extracellular ionic milieu.5 In vivo, cerebral ischemia has been postulated as one of several potential triggers for migraine attacks.6

Transcranial Doppler sonography may detect up to ≥10 microembolic signals per hour in patients with PFO and no apparent stenosis of the cerebrovascular bed.7 Most of these microembolic signals are thought to arise from thrombi of the venous circulation.8 In the vast majority, these clots silently embolize to the lung parenchyma and are lysed in the pulmonary circulation. In the presence of a PFO, paradoxical microembolism may ensue if right atrial exceeds left atrial pressure (e.g., after a Valsalva maneuver or while coughing). Assuming cerebral ischemia as trigger for migraine attacks in susceptible individuals and PFO as mediator for paradoxical embolism, the observed reduction in migraine frequency after percutaneous PFO closure appears pathophysiologically plausible. In addition, oral anticoagulation has also been reported to improve migraine, again suggesting a potential role for paradoxical embolism.9

Some limitations apply to the interpretation of our results. The incidence of tension-type headache and migraine declines with increasing age.10 Therefore, any reduction in the frequency of headache during follow-up may partially reflect the natural history. In addition, owing to the retrospective character of our survey, a recall bias of headache frequency cannot be excluded. However, four of five interventions were performed during the last 3 years. With a mean headache history of 24 years, it can be assumed that the period surveyed in the current study was not subjected to major recall bias. Furthermore, every patient was asked to indicate the ability to characterize headache frequency in relation to PFO closure and was excluded in case of insufficient memory. We did not assess in detail the use of headache medications before and after PFO closure. However, none of the patients needed supplemental migraine medication after PFO closure or went to see a headache specialist.

Patients were routinely treated with antiplatelet medication after PFO closure for the prevention of device-adherent thrombi for a duration of 6 months, until endothelialization of the implanted closure device was completed. Accordingly, only patients with a follow-up period of ≥12 months after PFO closure were included in the current study. The frequency of migraine attacks was assessed for the entire 12-month duration before and after PFO closure to reduce any confounding effect of antiplatelet therapy on migraine after the intervention.

PFO closure could also have a placebo effect on headache. However, neither at the time of percutaneous PFO closure nor during follow-up was any information of an association between migraine and PFO mentioned to the study participants. This and the lack of reduction of nonmigraine headaches render a sole placebo effect unlikely. Headache attacks in patients with migraine were reduced by >50%, whereas no reduction was observed in patients with nonmigraine headaches. However, before PFO closure can be recommended as a prophylaxis of migraine attacks, PFO closure will have to undergo a prospective trial.