Transcatheter patent foramen ovale closure mitigates aura migraine headaches abolishing spontaneous right-to-left shunting
Article Outline
Background
Patent foramen ovale (PFO) has been implicated in the etiology of migraine headache with aura (MHA), but the mechanisms that link right-to-left (R-to-L) shunt to MHA are unclear, and the reports on the efficacy of transcatheter PFO closure on MHA prevention are scarce.
Methods
We reviewed the clinical records of 131 consecutive patients who underwent successful transcatheter PFO closure at our institution at a mean age of 45 ± 13 years because of cryptogenic stroke. Of the 131 patients, 35 (27%) had a diagnosis of MHA made by the primary care physician or the referring neurologist. Migraine headache with aura incidence and severity were assessed by using Migraine Disability Assessment (MIDAS) questionnaire. Presence and magnitude of R-to-L shunt were assessed in all patients by means of transesophageal echocardiography and also by means of transcranial Doppler (TCD) in the last 50 patients (38%).
Results
Patients with MHA had a higher prevalence of thrombophilia (P = .007), a more complex atrial septal anatomy (P = .001), and they also had higher prevalence of spontaneous R-to-L shunt and of spontaneous large shunt, both at transesophageal echocardiography (P = .015, and .028, respectively) and at TCD (P = .036, and .038, respectively). After the procedure, 32 (91%) of 35 patients had either complete resolution or significant improvement in their MHA. At a mean follow-up of 1.7 ± 1.3 years, MHA disappeared completely in 29 (83%) of 35 patients. Of the remaining 6 patients, 3 patients (8%) had an improvement of ≥2 grades in the incidence and severity of MHA, 2 patients did not show any improvement of their MHA, whereas 1 patient reported a severe relapse of MHA about 1 year after the procedure.
Conclusions
In patients with PFO, MHA is associated with spontaneous large R-to-L shunt and thrombophilic conditions. Transcatheter defect closure seems to be an effective and safe means to treat MHA in patients with PFO.
Because of association studies, patent foramen ovale (PFO) has been implicated in the etiology of cryptogenic stroke secondary to paradoxical embolism,1, 2 neurological decompression illness in scuba divers,3, 4 and migraine headaches.5, 6 This association seems particularly high when PFO and migraine headaches with aura (MHAs) are considered, but the exact mechanisms that may explain this finding are not completely defined.7, 8
Few preliminary studies showed that a significant improvement in MHA after transcatheter PFO closure may occur,9, 10 but the mechanisms leading to MHA resolution are unknown.
The purpose of the present study was to examine the factors that may lead to MHA in patients with PFO and to assess the mechanisms through which PFO closure leads to an improvement of MHA.
Methods
Patient population
Between April 1999 and April 2005, a total of 131 consecutive patients with a PFO and at least one documented cryptogenic stroke underwent attempted transcatheter PFO closure at our institution using either the Amplatzer PFO device (AGA Medical, Minneapolis, MN) (n = 71), the CardioSEAL (NMT Medical Inc, Boston, MA) (n = 52), or the Helex device (WL Gore and Associates, Flagstaff, AZ) (n = 8). Age at PFO closure was 45 ± 13 years. Our baseline evaluation included a clinical examination and a screening for associated known cardiovascular risk factors and thrombophilia that were all performed before PFO closure, as reported.11
Patient selection
All patients enrolled in the present study fulfilled all the following criteria: (1) clinically and neuroradiologically confirmed ischemic stroke or symptoms unequivocally due to a transient ischemic attack with or without neuroradiologically identified intracranial ischemic lesions at cerebral computed tomography or magnetic resonance imaging; (2) exclusion of any identifiable cause for the thromboembolic event other than PFO by submitting all patients at least to the following investigations: cerebral computed tomography or magnetic resonance imaging, echo Doppler of the carotid and vertebral arteries, transesophageal echocardiography (TEE) with examination of the heart and thoracic aorta; and (3) presence of PFO associated with an atrial septal aneurysm (ASA) or PFO alone with a large (>30 microbubbles) spontaneous or induced right-to-left (R-to-L) shunt at contrast TEE within 3 cardiac cycles.
Procedure
The procedures were performed under general anesthesia, and venous access was gained from the right femoral vein. A complete right heart catheterization was performed with pressure recordings and blood sampling for oxymetry. The PFO was crossed with a multipurpose catheter, and the devices were implanted under fluoroscopy and TEE guidance as previously reported by others.12
Echocardiographic definitions
All echocardiographic evaluations were made with a multiplane TEE probe using agitated saline solution as a contrast agent. The criteria used to define an ASA at TEE were a base with a diameter of ≥15 mm and an excursion of the septum secundum of ≥10 mm.13 A PFO with R-to-L shunt (spontaneous or during Valsalva maneuver) was defined as the passage of at least 5 microbubbles from right to left atrium across the interatrial septum within 3 cardiac cycles from intravenous saline injection. The shunt through the PFO was defined as large when more than 30 microbubbles or cluster of bubbles were seen crossing the atrial septum at baseline or after Valsalva maneuver.14 Power M-mode Transcranial Doppler (TCD) scanning performed through the temporal approach during agitated saline injection was used as a complementary technique to assess the presence and degree of R-to-L shunt in the last 50 patients. Reported criteria were used to categorize shunt magnitude at TCD.15
Follow-up
After the procedure, medical treatment consisted of aspirin 100 mg/d for 12 months associated with ticlopidine 250 mg BID for the first 3 months in patients without thrombophilia, and of 6 months of warfarin therapy with a target international normalized ratio between 2 and 3 in all patients with thrombophilia. Closure status and magnitude of residual R-to-L shunt were evaluated serially with the following modality:
Assessment of MHA
Presence or absence of MHA was based on the assessment made by either the primary care physician or the referring neurologist. Migraine severity was assessed by using the validated Migraine Disability Assessment Questionnaire (MIDAS).16, 17 The MIDAS score was derived by adding up the number of functional days lost because of MHA and is graded 1 to 4.
Statistical analysis
Continuous variables are expressed as mean ± SD. Categorical variables are expressed as number (percentage). Statistical analysis was performed with Mann-Whitney test, Wilcoxon matched-pairs test, or Fisher exact test, when appropriate. Statistical significance was assumed at P < .05. GraphPad Prism PC 3.02 (GraphPad Software, San Diego, CA) was used for statistical calculations.
Results
Baseline characteristics of the study population are shown in Table I. Migraine headache with aura was present in 35 (27%) of 131 patients. Patients with MHA were more commonly females (P = .001), had a more complex atrial septal anatomy (P = .001), and had more often an associated thrombophilia (P = .007). The different types of thrombophilia found in patients with and without MHA are shown in Table II. Patients with MHA had a higher prevalence of spontaneous R-to-L shunt and a higher prevalence of spontaneous large shunt both at TEE (P = .015, and .028, respectively) and at TCD (P = .036 and .038, respectively). The prevalence of large R-to-L shunt during Valsalva was also higher in patients with MHA (P = .02). The prevalence and distribution of ischemic lesions at cerebral computed tomography and/or magnetic resonance imaging was similar in patients with and without MHA.
Table 1. Comparison of patients with PFO with versus without MHA
| MHA (n = 35) | No MHA (n = 96) | P | |
|---|---|---|---|
| Men | 10/35 (29%) | 58/96 (60%) | .001 |
| Cardiovascular risk factors | 27/35 (77%) | 68/96 (71%) | .520 |
| Thrombophilia | 19/35 (54%) | 27/96 (29%) | .007 |
| Atrial septal aneurysm | 26/35 (74%) | 41/96 (43%) | .001 |
| Ischemic lesions at CT or MRI | 27/35 (77%) | 63/96 (66%) | .290 |
| Multiple ischemic lesions at CT or MRI | 9/35 (26%) | 21/96 (22%) | .640 |
| Bilateral ischemic lesions at CT or MRI | 9/35 (26%) | 18/96 (19%) | .460 |
| Age at PFO closure (y) | 41.1 ± 11.0 | 47.0 ± 13.6 | .026 |
| R-to-L shunt at TEE | |||
 Spontaneous | 27/35 (77%) | 51/96 (53%) | .015 |
| Large spontaneous | 22/35 (63%) | 38/96 (40%) | .028 |
| Valsalva | 35/35 (100%) | 96/96 (100%) | 1.000 |
| Large Valsalva | 33/35 (94%) | 72/96 (75%) | .020 |
| Shunt at TCD | |||
| Spontaneous | 17/18 (94%) | 21/32 (66%) | .036 |
| Large spontaneous (curtain shower) | 13/18 (72%) | 12/32 (37%) | .038 |
| Valsalva | 18/18 (100%) | 32/32 (100%) | 1.000 |
| Large Valsalva (curtain shower) | 17/18 (94%) | 26/32 (81%) | .390 |
| Compliance rate | 0/35 (0%) | 1/96 (1%) | 1.000 |
| Residual shunt at TEE | |||
| Immediate | 11/35 (31%) | 23/96 (24%) | .500 |
| At follow-up | 6/35 (17%) | 12/96 (12%) | .570 |
| Length of follow-up (y) | 1.73 ± 1.36 | 1.95 ± 1.47 | .430 |
| Recurrent ischemic events | 0/35 (0%) | 4/96 (4%) | .570 |
Table 2. Types of thrombophilia found in patients with PFO with and without MHA
| Type of thrombophilia | MHA (n = 35) | No MHA (n = 96) | P |
|---|---|---|---|
| Hyperhomocysteinemia and MTHFRC677T mutation | 11/35 (31%) | 13/96 (13%) | .038 |
| Antiphospholipid antibodies | 4/35 (11%) | 3/96 (3%) | .082 |
| ProthrombinG20210A mutation | 2 (12%) | 4/96 (4%) | .658 |
| Factor VG1691A mutation | 2 (12%) | 4/96 (4%) | .658 |
| Other | 0 | 3/96 (3%) | .564 |
Patent foramen ovale closure
In 52 patients, PFO closure was achieved with a CardioSEAL device, in 71 patients with an Amplatzer PFO occluder, and a Helex device was used in 8 patients. The different types of device used in patients with and without MHA are shown in Table III. The devices were implanted successfully in all 131 patients. Because of malposition of the device across a wide ASA (diameter of the base = 32 mm), a patient underwent surgical removal of the device with the surrounding ASA and defect closure with autologous pericardium. One patient needed implant of 2 devices (CardioSEAL 28 and 23 mm) and another patient underwent percutaneous retrieval of an Amplatzer 25 mm and implantation of a larger Amplatzer device (35 mm) because of the presence of a multifenestrated PFO. Complication rate was similar in patients with and without MHA (0% vs 1%, P = .29). Immediate residual shunt was noted in 34 (26%; 11 patients with MHA and 23 without) of 131 patients, whereas a residual shunt was noted at 6 months in 18 (14%; 6 patients with MHA and 12 without) of 131 patients. Residual shunt in the follow-up was graded as trivial/mild in 14 patients (6 patients with MHA and 8 patients without). Four patients had a large residual shunt. Of the 4 patients, 3 underwent a successful re-do procedure to close the residual shunt.18 The fourth patient with large residual shunt refused re-do procedure.
Table 3. Devices implanted in patients with and without MHA
| Device type | MHA (n = 35) | No MHA (n = 96) | P |
|---|---|---|---|
| Amplatzer PFO occluder | 16/35 (46%) | 55/96 (57%) | .32 |
| CardioSEAL device | 15/35 (43%) | 37/96 (39%) | .69 |
| Helex device | 4/35 (11%) | 4/96 (4%) | .21 |
Migraine headache with aura
After the procedure, 32 patients (91%) had either complete resolution or significant improvement in their MHA. Table IV summarizes the effect of PFO closure in patients with known MHA. After a mean follow-up of 1.7 ± 1.3 years, MHA disappeared completely in 29 (83%) of 35 patients. Of the remaining 6 patients, 3 patients (9%) had an improvement of ≥2 grades in the incidence and severity of MHA as assessed by the MIDAS questionnaire. Three patients (all 3 without a residual shunt at TEE) did not improve after the procedure. In 2 patients, the incidence and severity of MHA remained unchanged, whereas 1 patient reported a severe relapse of MHA about 1 year after the procedure. No difference in the effect on MHA in patients treated with the Amplatzer versus CardioSEAL devices was seen.
Table 4. Pattern of MHA improvement in patients after PFO closure
| Before | After | P | |
|---|---|---|---|
| MHA episodes | |||
| Overall | 14.6 ± 23.2 | 1.7 ± 4.1 | .001 |
| Resolved completely | 10.4 ± 19.4 | 0 | .002 |
| MIDAS score | |||
| Overall | 21.5 ± 38.5 | 4.6 ± 8.2 | .02 |
| Resolved completely | 18.2 ± 43.4 | 0 | .01 |
Discussion
In patients with PFO, cryptogenic stroke is the result of paradoxical embolization of thrombus formed in the venous circulation carried by the R-to-L blood streaming across the interatrial septum, even if in situ formation of thrombus in the PFO tunnel has been proposed as an alternative mechanism. In support of the first of the proposed mechanisms for cryptogenic stroke in PFO, we were recently able to demonstrate that patients with PFO that have thrombophilia have a significantly increased risk of recurrent cerebral events.11
In patients with MHA on the other hand, the mechanisms that link R-to-L shunt to headache are unknown. It has been postulated that MHA is due to a small venous embolus that crosses the PFO paradoxically and passes to the cerebral circulation. Rather than inducing a stroke, the small embolus or platelet aggregate precipitates a spreading wave of depolarization that is recognized as the neurological phenomenon of migraine. Alternatively, it has been hypothesized19 that MHA is precipitated in susceptible individuals by chemical substances that can pass directly through the atrial shunt before they can be detoxified in their first passage through the lungs. This substance, in elevated concentration, could cause migraine in susceptible individuals without a PFO, but if a PFO is present it could potentially shunt from the venous to the arterial system and reach the brain in a more concentrated packet than if a central shunt was not present.
We observed in patients with PFO who experience MHA a significantly higher prevalence of spontaneous R-to-L shunt and a higher prevalence of large spontaneous R-to-L shunt both at TEE and at TCD. In patients with a PFO, cryptogenic stroke represents the consequence of a perfect spatial and temporal arrangement between 2 factors: formation of a thrombus of adequate size to occlude a relatively large-sized cerebral artery and its passage in near proximity of the interatrial septum; and incidental increase of right atrial pressure (Valsalva) so that R-to-L shunt can happen. This complicated arrangement may explain why cryptogenic stroke remains sometimes a rare event in patients with PFO.
The genesis of MHA in patients with PFO might be somehow different. Whatever theory of MHA is considered, a spontaneously large R-to-L shunt leads to a persistent or more frequent paradoxical shunting. Our findings support the hypothesis that in patients with PFO, persistent R-to-L shunting allows small-sized venous thrombus or platelet aggregate that would otherwise go undetected to enter the cerebral circulation, precipitating MHA.
Indeed, although the finding of a higher prevalence of large spontaneous R-to-L shunt in patients with PFO that have MHA could also stand for an increased paradoxical streaming of chemical factors, the significantly higher prevalence of thrombophilia seem to favor the idea of thrombus formation and subsequent paradoxical embolization as the leading cause of MHA in patients with PFO. Consistent with the hypothesis of embolization of small thrombi or platelet aggregates having a size insufficient to cause a clinically or radiologically evident cryptogenic stroke, we observed a similar prevalence of cerebral ischemic lesions in patients with PFO with and without MHA.
Another prominent finding of this observational study is that transcatheter PFO closure results in a profound improvement of MHA, with complete resolution in most of the patients (83%) and significant improvement in MHA incidence and severity (≥2 MIDAS grades) in a large proportion (50%) of the remaining patients. Of note, the significant improvement of MHA symptoms persisted for more than a year after the procedure. Our data suggest that the mechanism leading to MHA resolution is the abolishment of spontaneous R-to-L shunt rather than the abolishment of Valsalva-induced shunt. Indeed, we observed a long-lasting complete resolution of MHA also in the 6 patients who had trivial/mild residual R-to-L shunt induced by Valsalva maneuver. As far as thrombophilia is concerned, we think that the association between large spontaneous R-to-L shunt and increased tendency to thrombus formation may lead in some patients with a PFO to frequent formation and subsequent embolization of small emboli with a size unable to cause a stroke or an ischemic lesion, but able to trigger MHA.
Study limitations
This study is a retrospective evaluation and not a prospective randomized trial. The patient group being studied consisted of patients with a history of cryptogenic stroke with a known PFO and is not representative of the general population of patients with PFO. Although the MIDAS questionnaire is a well-established tool in the evaluation of the severity of MHA, it is a subjective method.
Conclusions
In patients with PFO, MHA is associated with spontaneous large R-to-L shunt and thrombophilic conditions. Paradoxical embolization of small-sized venous thrombus or platelet aggregate reasonably seems to be the leading mechanism precipitating MHA in patients with a PFO. Transcatheter defect closure seems to be an effective and safe means to treat MHA in patients with PFO.
References
- Patent foramen ovale as a risk factor for cryptogenic stroke. Ann Intern Med. 1992;117:461–465
- Clinical and imaging findings in cryptogenic stroke patients with and without patent foramen ovale: the PFO-ASA study. Stroke. 2002;33:706–711
- Patent foramen ovale: a review of associated conditions and the impact of physiological size. J Am Coll Cardiol. 2001;38:613–623
- Effect on migraine of closure of cardiac right-to-left shunts to prevent recurrence of decompression illness or stroke or for hemodynamic reasons. Lancet. 2000;356:1648–1651
- . Relationship between migraine and cardiac and pulmonary right-to-left shunts. Clin Sci. 2001;100:215–220
- Patent foramen ovale, a possible cause of symptomatic migraine: a study of 74 patients with acute cerebral ischemic stroke. Cerebrovasc Dis. 2002;13:102–106
- Potential source of cerebral embolism in migraine with aura: a transcranial Doppler study. Neurology. 1999;52:1622–1625
- Migraine with aura and right-to-left shunt on transcranial Doppler: a case-control study. Cerebrovasc Dis. 1998;8:327–330
- Percutaneous closure of patent foramen ovale reduces the frequency of migraine attacks. Neurology. 2004;62:1399–1401
- Migraine headache relief after transcatheter closure of patent foramen ovale. J Am Coll Cardiol. 2005;45:493–495
- Comparison of results of percutaneous closure of patent foramen ovale for paradoxical embolism in patients with versus without thrombophilia. Am J Cardiol. 2004;94:1012–1016
- Percutaneous closure of patent foramen ovale in patients with paradoxical embolism. Circulation. 2000;101:893–898
- Atrial septal aneurysm: a new classification in two hundred five adults. J Am Soc Echocardiogr. 1997;10:644–656
- Incidence of thrombus formation on the CardioSEAL and the Amplatzer interatrial closure devices. Am J Cardiol. 2004;93:426–431
- . Power M-mode Doppler for observing cerebral blood flow and tracking emboli. Ultrasound Med Biol. 2002;28:49–57
- Development and testing of the Migraine Disability Assessment (MIDAS) questionnaire to assess headache-related disability. Neurology. 2001;56:S20–S28
- Clinical utility of an instrument assessing migraine disability: the Migraine Assessment (MIDAS) questionnaire. Headache. 2001;41:854–861
- Transcatheter coil occlusion of residual shunt in patient with recurrent cryptogenic stroke after PFO percutaneous closure. Catheter Cardiovasc Interv. 2004;61:415–417
- Association of interatrial shunts and migraine headaches. J Am Coll Cardiol. 2005;45:489–492
PII: S0002-8703(05)00875-6
doi:10.1016/j.ahj.2005.09.019
© 2006 Mosby, Inc. All rights reserved.
