Abstract
Several publications have recently suggested that atrial septal defect may be associated with the physiopathology of headache with migraine-type characteristics. We describe four previously asymptomatic paediatric patients with atrial septal defect who underwent percutaneous Amplatzer septal occluder device implantation and who subsequently developed symptoms compatible with migraine headache. The cases had normal echocardiograms after the intervention and a benign course with headache improvement after several weeks or months. There are paediatric patients with atrial septal defect who may dramatically develop migraine symptoms with or without aura following percutaneous correction of their defect. Large paediatric studies are needed to offer accurate prognoses for children and their families. The possibility of using clopidogrel to treat this type of headache is subject to debate.
Introduction
Migraine is the most common type of childhood vascular headache. Its incidence and prevalence is high in statistical series (1). It can be classified into two main subtypes: with aura and without aura (2). Several genetic, vascular and neuronal factors have been identified as being involved in its pathogenesis.
Several publications (3–7) have recently suggested that certain alterations in the interatrial septum and their treatment by percutaneous catheterization and occlusion may be associated with the physiopathology of headaches with migraine-type characteristics: persistent foramen ovale (PFO) and atrial septal defect (ASD).
We describe four previously asymptomatic patients with ASD who underwent percutaneous Amplatzer septal occluder disc implantation in the Department of Paediatric Cardiology (Hospital de la Zarzuela) and who subsequently developed symptoms that were compatible with migraine headache. The sample consisted of 97 consecutive paediatric patients undergoing ASD closure with the Amplatzer septal occluder performed within a 24-month period. The procedures were performed under sedation (and not general anaesthesia) without contrast agent and were guided by transoesophageal echocardiography. The implantation technique has been detailed elsewhere (8). All patients received aspirin 50–500 mg/day 48–72 h before the procedure which was continued for 6 months.
Between 24 h and 7 days after closure, four of 97 patients developed intense headache. None of the other patients developed significant headache post ASD closure. All patients were followed up at 1 day, 1 week, 1 month, 3 months, 6 months and 12 months. During the interview they were specifically asked about the presence of headache.
When a migraine-type headache occurs de novo which is clearly related in time to an illness known to cause headache, the pain must be coded as a secondary headache according to the cause that produces it. This means that although we continuously refer to the concept of ‘migraine’, it is likely that the pain referred to should be coded differently once its precise physiopathology is determined, perhaps in section 6 (less probably in section 10) according to the criteria of the International Headache Society (IHS) (2).
Clinical cases
Case 1
A 16-year-old male with a negative family and personal history of headaches or other illnesses of interest was seen to have a pulmonary systolic murmur with fixed splitting of the second sound during the course of a routine check-up. Following transthoracic echocardiography, he was diagnosed with an ostium secundum ASD with significant left-to-right shunting (approximate Qp/Qs ratio of 1:9). It was seen to measure 20 mm by transoesophageal echocardiogram. The procedure was performed under plain anticoagulation with intravenous heparin (4000 IU). Complete transcatheter closure of the defect was performed (double Amplatzer occluder disc through a right venous femoral approach). No postoperative shunting was detected. The patient was treated with acetylsalicylic acid (ASA) at a dose of 300 mg/day. Forty-eight hours postimplantation, the patient began to notice progressive loss of vision in the left eye for 15 min accompanied by a diffuse, pulsating headache accompanied by photophobia, phonophobia, aggravation by routine physical activity and nausea without vomiting which lasted for approximately 3 h. Physical examinations following the crisis and the eye fundus were both normal.
As a result of these symptoms, cerebral magnetic resonance imaging (MRI) was performed (including diffusion-weighted), the results of which were normal. Two transthoracic echo-Doppler and one transcranial echo-Doppler failed to reveal any kind of alteration. The platelet count, platelet and coagulation activation tests were all normal. From that day onward he suffered pain of similar characteristics, albeit less intense and without aura every 2 or 3 days, lasting for 2 h, which gradually disappeared over a 2-month period until the patient became totally asymptomatic. Treatment consisted of ibuprofen during the crises. After 18 months of observation, the patient has not suffered headache again.
Case 2
A 13-year-old male with a negative family and personal history of headache or other illnesses of interest had an asymptomatic, right, parasternal systolic murmur, detected during a routine check-up. Consequently, a transthoracic sonogram was performed and he was diagnosed with an ostium secundum ASD with significant left-to-right shunting (Qp/Qs ratio of 2:1). It measured 25 mm by transoesophageal echocardiogram. Complete transcatheter closure of the defect was performed (Amplatzer septal occluder disc using a right venous femoral approach). The procedure was performed under plain anticoagulation with intravenous heparin (4000 IU). No postoperative shunting was detected. The patient was treated with ASA (200 mg/day). Twenty-four hours later, he developed photopsias and blurry vision, homonymous visual symptoms, paraesthesias and weakness in the upper left limb followed by a very intense, pulsating headache located in the right half of the head and accompanied by photophobia, phonophobia, aggravation by routine physical activity, nausea and vomiting. The episode lasted for 12 h; subsequently, he suffered similar headaches on a daily basis. Treatment with flunarizine was initiated 15 days following the onset of symptoms. The acute episodes did not improve despite treatment with ibuprofen or paracetamol. The frequency and intensity of the episodes decreased to one every 3 days, 3 weeks after the syndrome began. After 1 month with such a poor course of the symptoms, in light of the current state of knowledge and despite contraindications to its use in children, clopidrogel was proposed. However, it was agreed that this treatment would not be applied. Two cerebral MRIs, numerous transthoracic and transoesophageal echo-Dopplers and one transcranial echo-Doppler failed to reveal any kind of alteration. Platelet count and platelet and coagulation activation testing were all normal. Following 5 months of follow-up, the headaches persisted, albeit the intervals between them had become longer and longer (one every 4–5 days) and they tended to last for <1 or 2 h. In some of the crises, the patient took ibuprofen as soon as he noticed the aura (often an incomplete aura in comparison with the first episode) and the headache was thus aborted and the aura disappeared within the first hour. The patient continued to have periodic auras without headache for approximately 2 months, until he finally became completely asymptomatic with no relapses.
Case 3
A 14-year-old male with a negative family and personal history of headaches or other noteworthy conditions presented with a right, parasternal systolic murmur during a routine check-up. A transthoracic sonogram was performed and a 24-mm ostium secundum ASD with significant left-to-right shunting was diagnosed (Qp/Qs ratio of 2:1). After complete heparinization (4000 IU), transcatheter closure of the defect was performed (double Amplatzer septal occluder disc by percutaneous right femoral venous approach). The procedure was guided by transoesophageal echocardiography. No postoperative shunting was observed. Treatment with ASA was started 48 h before the percutaneous closure of the ASD (350 mg/day). On the third day he developed an excruciatingly severe, pulsating, diffuse headache with concurrent photophobia, phonophobia and aggravation by routine physical activity with associated nausea and vomiting. The pain lasted for 2–3 h, remitted spontaneously or following treatment with ibuprofen, and recurred every 2 days, with increasing latency between crises until they disappeared altogether at 3 months. The complementary testing failed to reveal any data of interest. Following a 1-year follow-up period, the headaches have not returned.
Case 4
During the course of a routine examination of a 4-year-old female with a negative family and personal history of headache or other remarkable illnesses a right, parasternal systolic murmur was discovered and a transthoracic sonogram subsequently performed. A 10-mm ostium secundum ASD with significant left-to-right shunting was diagnosed (approximate Qp/Qs ratio of 2:1). After complete heparinization, transcatheter closure of the defect was carried out (double Amplatzer septal occluder disc using a right venous femoral approach). No postoperative shunting was detected. Treatment with ASA was started 48 h before the procedure (60 mg/day). On the seventh day, she developed a headache that was described as pulsating and intense, diffuse, and accompanied by photophobia and phonophobia and intolerance to exercise without associated vomiting (doubtful nausea). The pain lasted for approximately 3 h and remitted spontaneously. It then reappeared every 2 or 3 days with a similar duration and spontaneously remitted or resolved with ibuprofen; the latency periods between crises increased until the headaches disappeared altogether after 4 weeks. Complementary testing, as in the other three cases, did not yield any data of interest. Following a 5-month follow-up, the patient is asymptomatic.
Discussion
It is important to distinguish between PFO (with right-to-left shunt) and ASD (with left-to-right shunt).
Persistent foramen ovale (PFO)
The PFO is a persistent connection between the left atrium and the right atrium, through which temporary right-to-left shunting can take place during Valsalva manoeuvres or if pulmonary hypertension develops (3, 9). Due to a valve mechanism, there is no left-to-right shunting despite the fact that it is favoured by the pressure difference between both atria. It is possible that the temporary right-to-left shunting has to do with the passing of venous microemboli into the systemic circulation (paradoxical embolism) (9–11). The relationship between paradoxical (right-to-left) embolism from a PFO and stroke remains controversial, despite numerous studies implicating this condition in the risk of stroke. In the SPARC study, PFO did not seem to be an independent risk factor for stroke in the general population (12–14). An association has also been established between migraine and PFO (7, 9), which is even greater in the case of migraine with aura (15). Primarily retrospective case–control studies have demonstrated a link between PFO closure and improvement of migraine headache. Few prospective data have confirmed the initial results. However, the only randomized, controlled trial finished to date analysing the effect of PFO closure on migraine failed to reach its primary outcome of resolution of migraine following the intervention. With currently ongoing trials, more information related to this topic can be expected (5, 16–18).
Atrial septal defect (ASD)
ASD causes a left-to-right connection. It is only at specific moments when there is a significant increase in venous return (Valsalva manoeuvre, respiratory variations with decreased intrathoracic pressure) or at the beginning of the left ventricular contraction that there might be a temporary, mild inversion of the shunting.
The possibility of ASD being a predictive risk factor for migraine is less evident (given that there have been no studies of sufficient statistical power) than in the case of PFO. Nevertheless, cases of exacerbated migraine with aura in prior migraine sufferers or de novo migraine have been described following percutaneous closure of some ASD (3), something that does not appear to hold true in the case of similar treatment procedures for PFO.
Studies of patients with headache following occlusion of ASD
Sharifi and coworkers reported 13 consecutive adult patients undergoing occlusion of ASD by percutaneous catheterization and occlusion using an Amplatzer device (4, 19). Between 8 and 40 h following closure, five of the 13 patients developed headache of sudden onset and significant intensity with migraine characteristics, three of them with and two without aura (all were treated with ASA). All the patients had a prior history of migraine, which in all cases was less severe than the current syndrome. They failed to respond to any of the drugs used (paracetamol, ibuprofen and sumatriptan). The authors suspected some kind of platelet derangement or microembolization and administered a loading dose of clopidogrel (300 mg). Within a mean time of 15 ± 8 min, the headache fully resolved in four patients and dramatically improved in the fifth. Treatment was continued at a dose of 75 mg/day for 1 month together with ASA for 6 months (325 mg/day) and there were no recurrences in the 9 months' follow-up. The authors considered that the most likely mechanisms underlying this situation were one or more of the following: (i) platelet activation; (ii) microembolization; (iii) secretion of platelet-released vasoactive substances. They recommend clopidogrel as concurrent treatment with ASA before and after closure of an atrial septal defect.
In Spain, Rodes-Cabau and coworkers (3) have described the case of a 31-year-old female with no prior history of headache who underwent a percutaneous ASD intervention (double Amplatzer septal occluder disc by percutaneous right femoral venous approach) treated with ASA. She developed photopsias and left scotoma 36 h later, followed by severe, right, hemicranial headache accompanied by photophobia, phonophobia, vomiting and paraesthesia in the left hand and foot that lasted for 24 h. She continued to have several similar episodes and migraine with aura was diagnosed according to IHS criteria (2). The frequency and intensity of the episodes diminished spontaneously without treatment over the following weeks. Cerebral computed tomography (CT), cerebral MRI, transcranial echo-Doppler and transoesophageal ecocardiogram were normal. After more than 3 months of evolution, the headaches disappeared. The results of numerous platelet counts and platelet and coagulation activation tests were normal.
Yankovsky and Kuritzky (6) reported the case of a 48-year-old male who suffered infrequent migraine attacks (once or twice per month) with visual and sensorial aura, who on the day following percutaneous closure of an asymptomatic ASD using an Amplatzer device developed a sensorial aura in the upper left limb followed by right visual hemianopsia, after which paraesthesia of the upper left extremity appeared with moderate headache. A transoesophageal echocardiogram and cranial CT were both normal. The headache became daily and was always preceded by some kind of aura; the frequency diminished only after 6 months of the procedure and the patient finally became asymptomatic.
Riederer and coworkers (20) reported the case of a 27-year-old female patient with a history of migraine with hemisensorial aura for the previous 14 years with a frequency of two crises per year, who, after undergoing intervention, suffered a significant change in her headaches. They became daily and lasted for longer. Various complementary tests were all normal, albeit the cerebral MRI performed a few days after the intervention revealed bilateral periventricular white-matter lesions. A control MRI scan with angiography sequences performed 3 weeks later showed no new lesions and no abnormalities of blood vessels. The authors believe that the lesions may have been caused by subclinical embolization during catheterization or even by the patient's migraine itself, and that the development of a thrombus at the site of the device is unlikely given that the echocardiographic check-ups were normal following implantation.
Mortelmans and coworkers (21) drew up a questionnaire about headache and administered it to 114 adult patients who had undergone percutaneous ASD closure using an Amplatzer device. The prevalence of migraine without aura fell from 19% to 12% following the procedure and prevalence of migraine with aura rose from 11% to 15% (P = 0.18 and P = 0.55, respectively). In 12 patients who were migraine sufferers prior to the percutaneous closure (four without and eight with aura) migraines disappeared. In this subgroup, the frequency of the migraine crises fell significantly (P = 0.01). De novo migraine was detected in 10 patients (seven with and three without aura). On the basis of these findings, the authors draw three main conclusions: (i) the percutaneous closure of the ASD is unrelated to a decrease in the prevalence of migraine; (ii) some migraine sufferers improve following ASD closure; (iii) a substantial number of patients who have never previously experienced migraine develop it for unknown reasons. These patients are relatively younger and bigger devices were used when compared with patients in whom migraine had disappeared.
Studies of PFO
Willmshurst and coworkers (5) demonstrated the phenomenon of an increase in the visual spectrum of migraine in 30% of patients during the immediate postoperative period following percutaneous PFO closure, although most of the patients had a prior history of migraine. Characteristic migraine pain is more common in patients with paradoxical cerebral embolism than in the general population (22). Many authors have documented significant migraine relief after PFO closure abolishing spontaneous right-to-left shunting (16, 22–27). Nevertheless, further large randomized trials are needed.
ASD closure has not been seen to be associated with remission of migraine with aura as definitively as in the case of PFO (23). Although it is possible that a subgroup of migraine sufferers with ASD also benefit from this closure, this has yet to be elucidated (21). It is important to point out that, as Tsimikas has stated (11), there are certain methodological limitations in these studies (scant distinction between the different types of atrial defects, small sample sizes, short follow-up times, the placebo effect, use of clopidogrel and prophylactic aspirin that are capable of decreasing the frequency of migraine on their own, and other factors).
What is the physiopathology?
Percutaneous closure of atrio-auricular defects is associated with a high success rate and is widely accepted as an alternative to surgery (3, 10). The most common complication is embolization in 1% of patients. Transient ischaemic attacks have been reported in 0.2–1% of cases during the first few days following the implantation of the device (3).
In most of the above-mentioned cases, PFO closure reduces the frequency of attacks in migraine sufferers and even aborts them entirely. In contrast to this, various hypotheses have been considered to account for the appearance of migraine symptoms subsequent to percutaneous ASD closure (left-to-right shunting). One possibility is cerebral thromboembolism through the cardiac cavities or due to the formation of thrombi on the surface of the device located on the atrial septum. Be that as it may, the possibility of paradoxical embolism is unlikely in cases in which a normal transcranial echo-Doppler and transoesophageal echocardiogram have been obtained following the intervention (3), as well as in those cases in which cerebral resonance is normal. Moreover, platelet and coagulation activation testing subsequent to implanting the device with normal results (absence of significant activation) suggests that the embolic mechanism or release of vasoactive substances from a thrombus that is adherent to the device is an implausible cause. It has been posited that the deformity or narrowing of the atrial septum caused by the device itself might induce the release of some vasoactive substances associated with migraine (3).
One hypothesis involves the function of atrial natriuretic protein in the physiopathology of headache in patients with ASD (with or without closure) (6). It is possible that this molecule could play a relevant role in the pathogenesis of migraine with aura (together with other molecules to which it is closely related, such as vasopressin), but it is not clear that the central nervous system secretion of atrial natriuretic protein is altered following correction of an ASD and that it has to do with the cause of the headaches (6, 28, 29), although it would be an interesting field of research.
Riederer (20) presents a case with periventricular lesions on cerebral MRI. The aetiology of different brain lesions in migraine sufferers is subject to debate (30). There is an increase in the prevalence of PFO and ischaemic brain lesions in individuals who have migraine with aura (23). Nevertheless, ASD correction should a priori increase cardiac output by eliminating the shunt from the systemic circuit to the pulmonary circuit, making cerebral ischaemia highly unlikely.
It is possible that there is an alteration in one or some physiological systems having a predominant circadian variation, which is the reason why the internal homeostatic adjustments diminish the impact of this variation after a few months (29). The most likely hypothesis is that of Wilmshurst and coworkers (31), based on serotoninergic hyperactivity (widely demonstrated and the basis for migraine treatment with triptans). Given that the lung is very rich in monoamino-oxidase, which degrades serotonin, it is logical that by restoring the septum (hence eliminating the left-to-right short circuit), there would be less serotoninergic clearance in the lungs. Serotonin is contained largely in the platelets, which, in turn, would be more amenable to aggregation in prone individuals.
Treatment
Wilmshurst and coworkers (31) have observed that when they added clopidogrel to treatment with ASA during the first month following percutaneous ASD closure there were fewer migraines with aura than with ASA alone [11 of 90 (12.2%) vs. 30 of 71 (42.3%), P < 0.001]. Furthermore, episodes of migraine were more severe and more frequent in patients who took only ASA, leading them to conclude that the combination of clopidogrel for 4 weeks and ASA for 6 months is superior to ASA alone for 6 months to prevent migraine with aura subsequent to percutaneous closure of an atrial shunt. This corroborates the hypothesis that involves platelet aggregation in the pathogenesis of the headaches.
Conclusions
The clinical cases in paediatric patients reported by our team and in the adult series reviewed have several characteristics in common: (i) a similar type of cardiac defect: ASD; (ii) the same transcutaneous catheterization technique using an Amplatzer occluder device; (iii) normal echocardiograms after the intervention; and (iv) benign course with improvement after several weeks or months.
Bearing these data in mind, we can conclude that: (i) there are paediatric patients with ASD who may dramatically develop migraine symptoms with or without aura following percutaneous correction of their defect; (ii) although prognosis for headache following occlusion of ASD in adults is usually good in the reviewed literature, large paediatric studies are needed to offer accurate prognoses for children and their families; (iii) the possibility of using clopidogrel to treat this type of headache is subject to debate. Nevertheless, the risks and safety of the drug have not been sufficiently studied and in the series published to date and in spite of the fact that the drug has generally been well tolerated, there have been cases of gastrointestinal (32) and subdural (33) haemorrhage, especially in combination with ASA; therefore, the indications are not yet entirely clear, bearing in mind that the clinical picture is benign.