The influence of percutaneous atrial septal defect closure on the occurrence of migraine

European Heart Journal (2005) 26, 1533 1537 doi:10.1093/eurheartj/ehi170 Clinical research The influence of percutaneous atrial septal defect closure on the occurrence of migraine Katrin Mortelmans 1, Martijn Post 1, Vincent Thijs 2, Luc Herroelen 2, and Werner Budts 1 * 1 Department of Cardiology, Internal Medicine, University Hospital Gasthuisberg, Herestraat 49, B-3000 Leuven, Belgium; and 2 Department of Neurology, University Hospital Gasthuisberg, Herestraat 49, B-3000 Leuven, Belgium Received 19 August 2004; revised 13 January 2005; accepted 20 January 2005; online publish-ahead-of-print 3 March 2005 This paper was guest edited by Per G. Bjornstad, Rikshospitalet, The National Hospital, Oslo, Norway KEYWORDS Migraine; Atrial septum defect; Transcatheter closure; Amplatzer device Introduction A causal relationship between the presence of migraine and a right-to-left shunt has been proposed. 1 3 Wilmshurst and Nightingale hypothesized that trigger substances in the venous circulation could pass through the interatrial septum and induce migraine attacks. These trigger substances should usually be filtered in the lungs, but in the presence of the right-to-left shunt the lungfilter may be bypassed. 4 This could explain the higher prevalence of a patent foramen ovale (PFO) in patients suffering from migraine. Moreover, it has been reported that closing an interatrial communication (mainly PFO) reduces migraine with aura and has a variable effect on migraine without aura. 2,5 8 Therefore, we were interested in determining the effect of percutaneous atrial septal defect (ASD) closure on the occurrence of migraine. * Corresponding author. Tel: þ32 16 344235; fax: þ32 16 344240. E-mail address: werner.budts@uz.kuleuven.ac.be Aims Percutaneous patent foramen ovale closure seems to influence migraine. We wanted to observe the effect of percutaneous atrial septal defect (ASD) closure on migraine. Methods and results All patients (.16 years of age) with a percutaneous ASD closure were selected from our database (n ¼ 114). A questionnaire about headache before and after closure was sent. According to the criteria of the International Headache Society, two neurologists diagnosed migraine with and without aura (MAþ and MA2, respectively). McNemar paired x 2 and Wilcoxon signed rank tests were used where applicable. Seventy-five patients (66%, 59 females, mean age 51 + 19 years) responded and were included in the study. An Amplatzer ASD occluder was used in all. Median follow-up time was 29 months (IQ1 and IQ3, 18 and 39 months, respectively). The prevalence of MA2 and MAþ changed from 19 (14/75) and 11% (8/75), respectively, before closure to 12 (9/75) and 15% (11/75), respectively, after closure (P ¼ 0.18 and P ¼ 0.55, respectively, vs. before closure). In 12 patients who suffered from migraine before closure (n ¼ 4 and 8, MAþ and MA2, respectively), migraine disappeared. In this subgroup, the frequency of migraine attacks decreased significantly (P ¼ 0.01). New-onset migraine was noted in 10 patients (n ¼ 7 and 3, MAþ and MA2, respectively). Conclusion Percutaneous ASD closure was not related to a decrease in prevalence of migraine. In a subgroup, patients who suffered from typical migraine before ASD closure, the frequency of migraine attacks decreased significantly. The reason for the new-onset migraine remains unexplained. A larger study sample will be necessary to determine these findings. Methods Patient selection All patients with a minimum age of 16 years, who underwent a percutaneous ASD closure in our centre between July 1999 and September 2003, were selected for the study. Patients were identified through the database of Congenital Cardiology, in which all percutaneous ASD closures of our hospital are registered. The medical files of these patients were reviewed with focus on the haemodynamic measurements and closing procedure of the ASD. The protocol of the study was approved by the institute s Ethics Committee, and all subjects gave informed consent. Evaluation of migraine A structured questionnaire was composed in collaboration with the Department of Neurology in a way such that a neurologist could diagnose migraine with or without aura, according to the International Headache Society. 9,10 The questionnaire was sent to all patients. They were asked about the presence, frequency, severity, duration, type, and site of headache. They were also interrogated about the occurrence of scotoma, paresthesia, paresis, aphasia, aggravating factors, accompanying symptoms such as nausea, vomiting, sono- and/or photophobia, and whether they where able to proceed daily activities. For the frequency of & The European Society of Cardiology 2005. All rights reserved. For Permissions, please e-mail: journals.permissions@oupjournals.org

1534 K. Mortelmans et al. migraine attacks, the relative frequency had to be reported by the patients (score from 0 to 7, no headache episodes to daily episodes, respectively, Figure 1). The questions focussed on three time periods: before, at 2 months, and at least 6 months after ASD closure. The patients were asked to send the questionnaire back, a prepaid envelope was enclosed. Two independent neurologists diagnosed migraine with and without aura (MAþ and MA2, respectively), before, at 2 months, and at least 6 months after closure, both blinded to the patients files and time periods. Statistical analysis Figure 1 Descriptive statistics was used to report patients and follow-up characteristics. Continuous variables with normal distribution were presented as mean + standard deviation. Median, first, and third quartiles (IQ1 and IQ3) were used when normal distribution was absent. McNemar paired x 2 test was performed to describe changes in proportions of patients with migraine. Kaplan Meier survival analysis was used to estimate the median time to migraine Translated example of the headache questionnaire. occurrence after ASD closure. Wilcoxon signed rank test was performed to compare migraine characteristics before and after ASD closure. One-way ANOVA and Tukey s honestly significant difference test were used to determine the differences between the haemodynamic parameters of the subgroups of patients. Interobserver reliability was evaluated by the Kappa coefficient. P, 0.05 was considered to be statistically significant. Results Patient characteristics One hundred and fourteen patients with ASD closure could be identified through the database of Congenital Cardiology. Seventy-five patients (66%, 59 females, mean age 51 + 19 years) sent the questionnaire back and could finally be included in the study. Mean body weight and mean length of the study population were 71 + 15 kg and 165 + 10 cm, respectively.

ASD and migraine 1535 Table 1 ASD characteristics in relationship with migraine before percutaneous closure Entire study population (n ¼ 75) No migraine (n ¼ 53) MAþ (n ¼ 8) MA2 (n ¼ 14) Age (years) 51 + 19 55 + 18 47 + 13 35 + 14 RA pressure (mmhg) 8 + 4 9+ 5 9+ 2 8+ 2 RV systolic pressure (mmhg) 29 + 10 30 + 11 22 + 2 29+ 6 RV diastolic pressure (mmhg) 8 + 4 8+ 4 7+ 2 8+ 2 PA systolic pressure (mmhg) 27 + 10 27 + 12 21 + 3 26+ 7 PA diastolic pressure (mmhg) 11 + 5 12+ 6 11+ 3 10+ 4 PA mean pressure (mmhg) 18 + 7 19+ 7 16+ 3 17+ 4 Q p /Q s (ratio) 1.9 + 0.4 1.9 + 0.4 1.9 + 0.6 2.1 + 0.5 Balloon sized defect size (mm) 23 + 7 22+ 7 23+ 7 24+ 6 Figure 2 RA, right atrium; RV, right ventricle; PA, pulmonary artery; Q p /Q s, ratio pulmonary cardiac output/systemic cardiac output. P ¼ 0.01 vs. no migraine (Tukey s post hoc test). Histogram of the different sizes of the Amplatzer ASD occluder. All ASDs where closed with the Amplatzer ASD occluder (AGA Medical Corporation, MN, USA). The haemodynamic measurements obtained during right heart catheterization are summarized in Table 1. The histogram of the different sizes of Amplatzer ASD occluders is given in Figure 2. Changes in prevalence of migraine Median follow-up time of the entire study was 29 months (IQ1 and IQ3, 18 and 39 months, respectively). Before ASD closure, typical MAþ was present in eight (8/75) and MA2 in 14 patients (14/75), a prevalence of 11 and 19%, respectively. Two months after closure, the prevalence of MAþ was changed into 15%. In four out of eight patients, who suffered from MAþ before closure, MAþ disappeared. However, seven patients who had no migraine before developed MAþ at 2 months after the procedure. In the same time period, the prevalence of MA2 changed into 13%. In six out of 14 patients who suffered from MA2 before closure, MA2 disappeared. However, two patients who had no migraine before developed MA2 at 2 months after the closing procedure. None of these changes in prevalence was statistically significant (P ¼ 0.55 and P ¼ 0.51 for MAþ and MA2, respectively, vs. before closure). The data for at least 6 months after the procedure were available for all patients, and the median time to the occurrence of migraine (MAþ or MA2) after ASD closure was 48 months (95% CI 41 55 months). The prevalence of MAþ remained 15% and no changes were reported between 2 Table 2 Changes in number of patients with migraine after ASD closure (n ¼ 75) Before closure After closure At 2 months No migraine 53 54 55 MA2 Persistent MA2 14 8 6 New-onset MA2 2 3 MAþ Persistent MAþ 8 4 4 New-onset MAþ 7 7 Values are number of patients. At least 6 months and 6 months. The prevalence of MA2 at least 6 months after closure was 12%. After at least 6 months, eight of 14 patients who had MA2 before closing procedure did not have MA2 anymore. However, three patients who did not have migraine before developed migraine in the time period of at least 6 months. None of these changes in prevalence was statistically significant (P ¼ 0.55 and P ¼ 0.18 for MAþ and MA2, respectively, vs. before closure). These data are summarized in Table 2. The interobserver reliability was high (kappa coefficient ¼ 0.7, P ¼ 0.045). Characteristics of migraine before and at least 6 months after ASD closure In the overall group of patients who suffered from migraine before ASD closure (MAþ and MA2, n ¼ 8 and 14, respectively), the median duration of migraine attacks in patients who still suffered from migraine at least 6 months after closure (MAþ and MA2, n ¼ 4 and 6, respectively) decreased from 270 min (IQ1 and IQ3, 90 and 2280 min, respectively) before to 120 min (IQ1 and IQ3, 60 and 270 min, respectively) at least 6 months after ASD closure (P ¼ 0.24). In these patients, the severity of migraine attacks remained unchanged. The median severity score (on a scale ranging from 0 to 10, minimal to very severe pain, respectively) was 2 (IQ1 and IQ3, 1 and 4, respectively) before and after ASD closure (P ¼ 0.41). The median scores

1536 K. Mortelmans et al. indicating the frequency of migraine attacks in the entire group of patients who suffered from migraine before ASD closure were 3 (IQ1 and IQ3, 2 and 4, respectively) before and 3 (IQ1 and IQ3, 0 and 4, respectively) at least 6 months after ASD closure. Although the median scores were similar, Wilcoxon signed rank test was statistically significant (P ¼ 0.01). The latter score was calculated from a scale ranging from 0 to 7 (from no headache to headache s several times per day). Relationship between ASD characteristics before closure and migraine Table 1 summarizes the haemodynamic characteristics of the ASDs before closure. The entire study population is divided in three subgroups (patients without migraine, MAþ, and MA2). No significant differences could be identified among the three subgroups, except for MA2 patients who were significantly younger than patients with no migraine (P ¼ 0.01). MAþ patients were also younger than patients with no migraine, but significant statistical difference was not reached. Table 3 summarizes the haemodynamic characteristics of the ASDs after closure. For this purpose, the entire study population was split into four subgroups (patients in whom no migraine was present during the study, patients who developed migraine after ASD closure, patients in whom migraine disappeared after percutaneous intervention, and patients in whom migraine persisted). Patients in whom no migraine was present during the entire study were significantly older than patients in whom migraine persisted or developed (P ¼ 0.0001 and P ¼ 0.045, respectively). In addition, the devices used for closure were significantly smaller in patients in whom no migraine was present during the entire study and in whom migraine disappeared when compared with the sizes of devices used in patients in whom migraine developed and persisted, respectively (P ¼ 0.01 and P ¼ 0.004, respectively). Discussion We found in patients with a secundum type ASD a prevalence of migraine of 30%. Eleven per cent suffered from MAþ and Table 3 ASD characteristics in relationship with migraine after percutaneous closure No migraine before and after (n ¼ 43) 19% from MA2. After percutaneous closure with the Amplatzer ASD occluder, the overall prevalence of migraine remained unchanged. However, we found that the frequency of migraine attacks in patients with migraine before closure decreased significantly for a mean follow-up period of at least 2 years. In addition, a substantial number of patients developed migraine after ASD closure. These patients were relatively younger and bigger devices were used when compared with patients in whom migraine disappeared. The prevalence of migraine in the overall population is 6% in men and 15 18% in women. 11 14 A higher prevalence of PFO is reported in patients with migraine when compared with the general population, 36 38 vs. 25%. 1 3 In addition, we reported recently a higher prevalence of migraine in patients with a PFO. 7 All this suggests a causal relationship between the presence of PFO and migraine. Some authors hypothesize that a particular genetic substrate might determine both atrial septum abnormalities and migraine. 15 However, migraine is a form of neurovascular headache, in which not only neural events but also vasoactive trigger substances might play an important role. 16 Wilmshurst and Nightingale 4 suggested that trigger substances in the venous circulation could pass through a PFO. Normally, these trigger substances will be filtered in the lungs. In individuals with a right-to-left shunt, a lower dose of trigger substances is needed to induce migraine attacks because the shunt permits the lungfilter to be bypassed. In individuals without a right-to-left shunt, a larger amount of trigger substances is needed to induce migraine by overwhelming the filter capacity of the lungs. Micro-emboli and vasoactive chemicals such as serotonin are proposed as potential trigger substances. This hypothesis might explain why recent studies report a decrease in prevalence and the severity of migraine after PFO closure. 7,8 Until now no isolated data have been published about the relationship between migraine and the presence of an ASD. Although an ASD is mainly characterized by a left-to-right shunt and volume overload of the right heart, a small right-to-left shunt may occur during Valsalva or exercise. Theoretically, this right-to-left shunt could also permit that the lungfilter is bypassed. Indeed, we found a relatively high prevalence of migraine in patients with an unclosed secundum type ASD (up to 30%). This seems to be higher Migraine persisted (n ¼ 10) Migraine developed (n ¼ 10) Migraine disappeared (n ¼ 12) Age (years) 58 + 17 35 + 12 29 + 18 52 + 13 RA (mmhg) 9 + 5 7+ 2 5+ 4 9+ 2 RV systolic pressure (mmhg) 30 + 2 26+ 3 23+ 7 29+ 4 RV diastolic pressure (mmhg) 8 + 4 7+ 2 6+ 3 9+ 2 PA systolic pressure (mmhg) 28 + 12 24 + 4 20+ 6 27+ 9 PA diastolic pressure (mmhg) 12 + 6 9+ 3 10+ 3 12+ 4 PA means pressure (mmhg) 19 + 8 16+ 3 13+ 4 18+ 5 Q p /Q s (ratio) 1.9 + 0.4 2.1 + 0.5 1.9 + 0.3 1.8 + 0.5 Defect size (mm) 21 + 6 26 + 5 } 25 + 7 17+ 4 RA, right atrium; RV, right ventricle; PA, pulmonary artery; Q p /Q s, ratio pulmonary cardiac output/systemic cardiac output. P ¼ 0.0001 vs. migraine persisted. P ¼ 0.005 vs. migraine developed. P ¼ 0.017 vs. migraine persisted. } P ¼ 0.004 vs. migraine disappeared (Tukey s post hoc test).

ASD and migraine 1537 than in the general population, but lower than in patients with a PFO. 7 The latter can be explained in different ways. First, in patients with a PFO, the right-to-left shunt is obligate; whereas in patients with an ASD, the left-toright shunt predominates and permits less vaso-active agents passing through the interatrial septum. Indeed, in almost all studies where the occurrence of migraine is influenced by percutaneous closure, a clinical event related to a obligate right-to-left shunt was present. 7,8 Secondly, the age distribution of our ASD population differed slightly from PFO populations 7,8 and could explain differences in prevalence of migraine. We found that ASD patients with migraine were younger than patients with no migraine, and that MAþ were older than MA2 patients. The latter may support the hypothesis that MAþ in older patients (more prone for thrombus formation in the venous circulation) could be related with a right-to-left shunt through an interatrial communication. Indeed, in a subgroup analysis after ASD closure, migraine disappeared only in the relatively older patient (.50 years of age). In contrast to PFO closure, we found no significant changes in the overall prevalence of MAþ or MA2 after ASD closure. Nevertheless, in several patients MAþ or MA2 disappeared after closure, but in an almost equal number of patients MAþ and MA2 developed. However, in the subgroup of patients with migraine before closure, the frequency of migraine attacks decreased and remained decreased for a median follow-up time of at least 2 years. The latter excludes the influence of drugs or a possible placebo effect. The reason why patients developed migraine after percutaneous ASD closure remains unexplained. Yankovsky and Kuritzky 17 reported aggravation of MAþ into a daily pattern after ASD closure and suggested that elevated or changed levels of atrial natriuretic peptide as a consequence of intra-atrial pressure imbalance immediately after closure could play an important role for the explanation of their findings. Others believe more in a disturbed circadian variation of a basic physiologic process. 18 Moreover, it is not excluded that on the left sided disk of the Amplatzer device micro-thrombi may be formed and micro-embolize in the systemic circulation. The finding that relatively bigger devices were used in patients who developed migraine or in whom migraine persisted when compared with patients in whom migraine disappeared, may support this hypothesis. However, the explanation of the changes in incidence of migraine remains speculative in the absence of large prospective randomized trials. Finally, we want to apply a few limitations to our study. First, a selection bias might be present because our patient population was a selected cohort referred to our centre for percutaneous ASD closure and 34% of the selected patients did not send the questionnaire back. Secondly, it was a retrospective study in which patients were asked about their headache before and after ASD closure. It might be difficult for the patients to remember the characteristics of headache before and after ASD closure. Therefore, a recall bias might be introduced. Thirdly, the study was underpowered to detect significant differences in MA2 and MAþ characteristics. In conclusion, percutaneous ASD closure was not related to a decrease in prevalence of migraine. However, in the subgroup patients who suffered from typical migraine before ASD closure, the frequency of migraine attacks decreased significantly. These data fit with the hypothesis that vaso-active agents passing through the interatrial septum might play a role in the occurrence of migraine. The reason for the new-onset migraine remains unexplained. A larger study sample will be necessary to confirm and analyse these findings. References 1. Del Sette M, Angeli S, Leandri M, Ferriero G, Bruzzone GL, Finocchi C, Gandolfo C. Migraine with aura and right-to-left shunt on transcranial Doppler: a case-control study. Cerebrovasc Dis 1998;8:327 330. 2. Sztajzel R, Genoud D, Roth S, Mermillod B, Le Floch-Rohr J. Patent foramen ovale, a possible cause of symptomatic migraine: a study of 74 patients with acute ischemic stroke. Cerebrovasc Dis 2002;13:102 106. 3. Anzola GP. Clinical impact of patent foramen ovale diagnosis with transcranial Doppler. Eur J Ultrasound 2002;16:11 20. 4. Wilmshurst P, Nightingale S. Relationship between migraine and cardiac and pulmonary right-to-left shunts. Clin Sci (Lond) 2001;100:215 220. 5. Wilmshurst PT, Nightingale S, Walsh KP, Morrison WL. Effect on migraine of closure of cardiac right-to-left shunts to prevent recurrence of decompression illness or stroke or for haemodynamic reasons. Lancet 2000;356:1648 1651. 6. Morandi E, Anzola GP, Angeli S, Melzi G, Onorato E. Transcatheter closure of patent foramen ovale: a new migraine treatment? J Interv Cardiol 2003;16:39 42. 7. Post MC, Thijs V, Herroelen L, Budts WI. Closure of a patent foramen ovale is associated with a decrease in prevalence of migraine. Neurology 2004;62:1439 1440. 8. Schwerzmann M, Wiher S, Nedeltchev K, Mattle HP, Wahl A, Seiler C, Meier B, Windecker S. Percutaneous closure of patent foramen ovale reduces the frequency of migraine attacks. Neurology 2004;62: 1399 1401. 9. Headache Classification Committee of the International Headache Society. Classification and diagnostic criteria for headache disorders, cranial neuralgias and facial pain. Cephalalgia 1988;8:1 96. 10. Goadsby PJ, Olesen J. Diagnosis and management of migraine. BMJ 1996;312:1279 1283. 11. Stewart WF, Shechter A, Rasmussen BK. Migraine prevalence. A review of population-based studies. Neurology 1994;44:S17 S23. 12. Lusic I. Population variation in migraine prevalence the unsolved problem. Coll Antropol 2001;25:695 701. 13. Lipton RB, Stewart WF, Reed M, Diamond S. Migraine s impact today. Burden of illness, patterns of care. Postgrad Med 2001;109:38 40, 43 45. 14. Steiner TJ, Scher AI, Stewart WF, Kolodner K, Liberman J, Lipton RB. The prevalence and disability burden of adult migraine in England and their relationships to age, gender and ethnicity. Cephalalgia 2003;23:519 527. 15. Arquizan C, Coste J, Touboul PJ, Mas JL. Is patent foramen ovale a family trait? A transcranial Doppler sonographic study. Stroke 2001; 32:1563 1566. 16. Goadsby PJ, Lipton RB, Ferrari MD. Migraine current understanding and treatment. N Engl J Med 2002;346:257 270. 17. Yankovsky AE, Kuritzky A. Transformation into daily migraine with aura following transcutaneous atrial septal defect closure. Headache 2003;43:496 498. 18. Gupta V. Closure of atrial septal defect and migraine. Headache 2004;44:291.