Migraine Intervention With STARFlex Technology (MIST) Trial
A Prospective, Multicenter, Double-Blind, Sham-Controlled Trial to Evaluate the Effectiveness of Patent Foramen Ovale Closure With STARFlex Septal Repair Implant to Resolve Refractory Migraine Headache
Background— Patent foramen ovale (PFO) is prevalent in patients with migraine with aura. Observational studies show that PFO closure resulted in migraine cessation or improvement in ≈80% of such patients. We investigated the effects of PFO closure for migraine in a randomized, double-blind, sham-controlled trial.
Methods and Results— Patients who suffered from migraine with aura, experienced frequent migraine attacks, had previously failed ≥2 classes of prophylactic treatments, and had moderate or large right-to-left shunts consistent with the presence of a PFO were randomized to transcatheter PFO closure with the STARFlex implant or to a sham procedure. Patients were followed up for 6 months. The primary efficacy end point was cessation of migraine headache 91 to 180 days after the procedure. In total, 163 of 432 patients (38%) had right-to-left shunts consistent with a moderate or large PFO. One hundred forty-seven patients were randomized. No significant difference was observed in the primary end point of migraine headache cessation between implant and sham groups (3 of 74 versus 3 of 73, respectively; P=0.51). Secondary end points also were not achieved. On exploratory analysis, excluding 2 outliers, the implant group demonstrated a greater reduction in total migraine headache days (P=0.027). As expected, the implant arm experienced more procedural serious adverse events. All events were transient.
Conclusions— This trial confirmed the high prevalence of right-to-left shunts in patients with migraine with aura. Although no significant effect was found for primary or secondary end points, the exploratory analysis supports further investigation. The robust design of this study has served as the model for larger trials that are currently underway in the United States and Europe.
Received July 12, 2007; accepted December 6, 2007.
Migraine affects ≈13% of the general population between 20 and 64 years of age1 with a male-to-female ratio of 1:3, and in ≈36% of patients, the attack is preceded by an aura.2 Migraine with aura is associated with patent foramen ovale (PFO), a remnant of the fetal anatomy, and with other causes of right-to-left shunts (RLSs).3 In patients with significant PFOs and/or RLSs, the prevalence of migraine with aura is increased.4 In cadaver5 and live population studies,6 total PFO prevalence is reported at 27%, of which 4.9% were large at rest and an additional 2.4% were on Valsalva maneuver. It has been postulated that in some migraine patients, venous blood contains agents normally removed by passage through the lungs that can trigger an attack of migraine if they reach the brain in sufficient concentrations; alternatively, long-term shunting of the agents may reduce the threshold for migraine generation in the brain.7
Editorial p 1358
Clinical Perspective p 1404
About 80% of patients who underwent PFO closure for nonmigraine indications reported cessation or improvement in their migraine attacks after PFO closure.6,8–10 These studies are limited by being predominantly retrospective, nonrandomized, and conducted in highly selected populations of patients. Furthermore, the highly variable course of migraine and the known placebo effect in previous migraine trials11 mean that for proper conclusions to be drawn, a controlled and blinded trial design is imperative. The Migraine Intervention With STARFlex Technology (MIST) trial was a randomized controlled study designed to assess the effect of PFO closure on migraine headache in patients with frequent, disabling, and drug-resistant migraine with aura.
This was a prospective, multicenter, randomized, double-blind, sham-controlled clinical trial. The study was approved by a multicenter research ethics committee in the United Kingdom. Patients gave written informed consent at each of the 3 stages of the screening process (at medical screening with a headache specialist, at the cardiology contrast echocardiography visit, and before randomization at an implantation center). All procedures were conducted in accordance with the most recent revision (2004) of the Declaration of Helsinki.
Patients were identified from records of participating headache centers or by self-referral after preliminary screening on a Web site (www.migraine-mist.org). They were offered a headache specialist screening visit if they were 18 to 60 years of age with a history of migraine with aura as defined by the criteria of the International Headache Society12 starting before 50 years of age; had ≥5 migraine headache days per month but at least 7 headache-free days per month; and reported a history of having failed at least 2 classes (β-blockers, anticonvulsants, calcium channel blockers, tricyclics, and serotonin antagonists) of preventive medication because of inefficacy or intolerability as judged by an investigator. Patient history at the first visit to the headache doctor determined the inclusion criteria, and patients were not excluded after the minimum 30-day baseline period if the number of headache days fell outside the inclusion criteria.
Exclusion criteria included other cardiovascular defects, the presence of intracardiac thrombi, active endocarditis, coagulopathy, bacteremia or active infections, elevated serum creatinine, platelet disorder, other neurological disorders, recent history of active peptic ulcer or gastrointestinal bleeding, cirrhosis, portal hypertension or pulmonary arteriovenous malformation, contraindication to aspirin or clopidogrel, or any other medical condition or contraindication to the procedures and treatments used in the study. Patients also were excluded if they were pregnant, were planning pregnancy, or were nursing over the duration of the study, as were those who required PFO closure for reasons unrelated to migraine, ie, stroke or decompression illness. The patient flow through the study is illustrated in Figure 1.
Eligible patients were referred for assessment for the presence of an RLS by contrast transthoracic echocardiography at 1 of 2 echocardiography laboratories according to a specified protocol. This diagnostic method was selected because it combined a low-risk approach compared with transesophageal echocardiography with a high degree of sensitivity and specificity.13 Images were acquired in the apical 4-chamber view using second harmonic imaging. Agitated saline was injected into an antecubital vein at rest and during provocative maneuvers (Valsalva maneuver, sniff, cough). Shunt size was determined using a practical, clinical hybrid method based on approximate count and visual appearance of bubbles in the left heart during the first 5 cardiac cycles of contrast entering the right atrium. See the online-only data supplement for more details. Patients with small or no shunts were excluded and referred back to the headache specialist for further care. Patients with evidence of a moderate or large RLS, interpreted as resulting from the presence of a PFO, were referred to a trained and experienced (minimum of 10 PFO closure procedures) interventional cardiology center for randomization. Headache diaries were recorded for at least 30 days before randomization (baseline phase).
Randomization: PFO Closure Procedure or Sham Procedure
Aspirin and clopidogrel were given to all patients as a loading dose in the 24 hours before the procedure (300 mg each) and for 90 days after the procedure (75 mg each daily). After induction of general anesthesia, all patients underwent transesophageal echocardiographic assessment of the interatrial septal anatomy to ensure that no anatomic contraindication to PFO closure was present. The patient was then randomized by the investigator who telephoned a central computerized service. Patients were randomized in a 1:1 ratio (blocks of 4) to either PFO closure with the STARFlex septal repair implant (NMT Medical Inc, Boston, Mass) or a sham procedure (skin incision in the groin). Patients randomized to implant were given intravenous heparin 100 IU/kg periprocedurally as required to keep activated clotting time >200 seconds. Only the staff present in the cardiac catheterization laboratory knew the treatment allocation. All patients were subsequently managed in an identical fashion and were reviewed before discharge. Patients and headache specialists were not informed of treatment allocation during follow-up.
Patients attended headache clinics after the procedure for 6 visits at intervals of 30±7 days. Days 0 to 90 were defined as the healing phase; days 91 to 180, as the analysis phase. During this time, patients were encouraged to continue with existing migraine prophylactic medications and not to initiate new medications. Patients were allowed to use rescue medications at any time to treat migraine attacks. A final study visit was conducted by the implanting cardiologist, who informed the patient of his or her treatment allocation and assessed the implant arm for residual shunts by repeat transthoracic echocardiography.
Daily headache diaries were kept, and at each clinic visit, patients completed the Headache Impact test (HIT-6)14 and the Short-Form 36 (SF-36v2) Quality of Life questionnaire.15 At baseline, the end of the healing phase, and the end of the analysis phase, patients completed the Migraine Disability Assessment (MIDAS) questionnaire.16
Primary Efficacy End Point
The primary efficacy end point was migraine headache cessation during the analysis phase. It was derived from diary data.
Secondary Efficacy End Points
Secondary efficacy comparisons were incidence of migraine during the healing phase; change in the severity of migraine attacks based on MIDAS (over a 3-month retrospective period) and HIT-6 (over a 1-month retrospective period) scores; change in the frequency of migraine attacks other than elimination of attacks; change in the characteristics of migraine (with or without aura and change thereof); change in the severity, frequency, and character of migraine relative to effective closure rate or presence of residual leak; and change in quality of life based on the SF-36v2 questionnaire (over a 1-month retrospective period).
Unless indicated otherwise, secondary efficacy comparisons were of the change between the baseline and analysis phases. The estimation of total migraine headache days was defined as the number of migraine headaches times the average length of the migraine in hours divided by 24 and rounded up to the nearest day.
Secondary Safety End Points
Adverse events were recorded at all clinic visits. Prespecified safety end points included device and procedural success and the incidence of major adverse events, including death, stroke, bleeding complications, and adverse drug reactions. Adverse events were monitored by a data, safety and adverse events monitoring board (DSAEMB) that included 4 physicians (3 cardiologists and 1 neurologist), a medical ethicist, and a biostatistician who were independent of the trial investigators.
All randomized patients formed the intention-to-treat population, which was the population for the primary analyses of efficacy and safety. Efficacy analyses also were conducted on a per-protocol population, defined as all randomized patients who received the allocated treatment and who had completed follow-up.
On the basis of previous observational studies,6,8,9 we anticipated cessation of migraine in 40% of the implant group compared with 15% of the sham group. A sample size of 132 patients was required for 80% power using a 2-sided test with P=0.05. Allowing for a 10% dropout rate and a further 4% loss of blinding for medical reasons, we aimed to randomize 150 patients.
All significance testing between the 2 groups was 2 sided and performed at P=0.05, with no adjustment for multiple comparisons. The primary efficacy end point was analyzed with Fisher exact test because of the low incidences involved. Secondary end points were analyzed with the χ2 test if the data were dichotomous (eg, migraine incidence and device success) or by the Wilcoxon rank-sum test if they were continuous (eg, attack frequency). Adverse event frequency was compared with the χ2 test.
The study was funded by NMT Medical Inc and designed jointly by NMT Medical Inc and a scientific advisory board (the MIST Trial Design Physician Advisory Group), together with additional advisors on bioethics, biostatistics, and patient groups. The study was managed by a steering committee and the DSAEMB.
The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.
Patient flow through the trial is shown in Figure 2. A total of 432 patients were assessed for an RLS by transthoracic echocardiography. The types of RLS detected by the procedure are shown in Table 1. A shunt was detected in 260 patients (60%), of which 163 (38%) were interpreted as being due to a moderate or large PFO. Of the patients with other shunts, 96 (22.2%) had small shunts or large pulmonary shunts and 1 (0.2%) ASD. The baseline characteristics of the 147 patients (16 patients did not progress to randomization: 6 because of personal reasons or lost to follow up, 6 for medical reasons including pregnancy, dental treatment, sinusitis, hysterectomy, steroid treatment and late declaration of aspirin sensitivity, and 4 others after transoesophageal echocardiography; 2 were diagnosed as having an ASD and it was not possible to confirm PFO in the other 2) who subsequently underwent randomization (74 to PFO closure, 73 to sham) are given in Table 2. The 2 groups were well matched in terms of age, gender, and race. At baseline, patients also were similarly matched in terms of the average frequency of migraine attacks, headache impact (MIDAS and HIT-6 scores), and median number of acute and preventive medications being taken.
No PFO was found or crossed in 5 of the 74 patients (7%) randomized to closure. In 1 patient, a 23 mm device embolised to right atrium after release and in a second patient, the initial implant position was unsatisfactory, with prolapse of left atrial arms into the right atrium. This device was withdrawn from the PFO but subsequently embolised to the left pulmonary artery whilst being withdrawn into the delivery sheath. Both devices were successfully retrieved using snares. In a third patient, the initial implant could not be deployed and was retrieved without being detached. All 3 patients had a second device successfully implanted and continued in the study. One randomized patient was withdrawn because of procedure-related cardiac tamponade before device deployment. Two patients in each group withdrew as a result of adverse events in the follow-up period. One patient was withdrawn after being lost to follow-up. Therefore, the study population consisted of 147 patients in the intention-to-treat and 136 in the per-protocol analyses.
The major efficacy analyses are presented for both the intention-to-treat and per-protocol populations in Tables 3 and 4⇓. The primary end point of migraine cessation was observed for 3 patients in each group. Secondary end points did not differ significantly between groups for either the intention-to-treat or per-protocol populations.
Recognizing the failure to achieve predefined endpoints, we conducted exploratory analysis17 to aid hypothesis generation and future study design. Two patients in the implant group were noted to account for 20% of all headache days in the implant group during the analysis period (Figure 3) and differed from the rest of the population (Shapiro-Wilk test, P=0.0014). When these patients were excluded from the per-protocol population, a significant 2.2 d/mo (from 6.0 to 3.8 d/mo; 37%) reduction was noted in median total migraine headache days for the implant group compared with 1.3 d/mo (from 5.0 to 3.7 d/mo; 26%) in the sham group (P=0.027).
Residual moderate or large atrial level shunts were reported in 4 patients when assessed at 6 months by the treating cardiologists, with no differences seen in treatment effect between those closed versus those with a residual shunt. No significant changes could be observed in the severity end points of the MIDAS or HIT-6 scales or in the quality of life end point SF-36v2.
Tolerability and Safety
Most patients in both groups reported ≥1 minor adverse events, most commonly attributed to trial antiplatelet medication. Serious adverse events occurred in 16 patients (Table 5). Other procedural complications included pericardial effusion in 2 patients, 1 of which required percutaneous drainage, and a retroperitoneal bleed in 1 patient in the implant group, which was managed conservatively. Patients in the sham group experienced 3 serious adverse events that were probably related to antiplatelet medication (incision site bleed, anemia, and nosebleed). The patient in the sham arm who suffered a stroke 4 months after the procedure and 1 month after withdrawal of antiplatelet medication was withdrawn and later underwent PFO closure. In 3 patients, devices were withdrawn due to dissatisfaction with the initial implant position. A second device was deployed in a satisfactory position during the same procedure in all 3 patients.
The premise of closure of PFO to reduce migraine frequency continues to be researched9,18,19; however, the MIST trial is the first prospective, randomized, placebo (sham) -controlled trial of PFO closure for the treatment of migraine with aura. The lack of objective measures of migraine and the known placebo effect seen in previous pharmacological studies20 meant that adequate blinding of both patients and headache physicians was an important element in the design of the MIST trial. Although not assessed formally, we believe blinding was achieved with the sham procedure. We have demonstrated that a sham procedure is feasible in a device trial and recommend that it become the standard for future trials of PFO closure for migraine. Placebo control can be problematic in surgical procedures, but a sham procedure has been used in 3 controlled studies of acupuncture for migraine and led to valid results.21–23
The criteria for patient selection included only migraine with aura patients with frequent and refractory attacks. Our results demonstrated that the study patients had ≈5 migraine attacks in the month before treatment (diary), with ≈30 days of headache in the previous 3 months (MIDAS). The baseline MIDAS score was 36 and the HIT-6 score was 67, both in the range of severe headache impact.14,16 It should be noted that it is possible for patients with >5 migraine headache days per month but effective acute/rescue medications to score low on MIDAS because the score is calculated by adding time lost and time at <50% of normal capability in daily activities.16 The patient selection criteria were therefore met in the study population, which was well matched between the 2 groups. In general, patients were taking few prophylactic medications at baseline, supporting the suggestion of relative failure of these treatments in the past (entry criteria was failure of ≥2 classes of prophylactic medications). However, on average, patients were taking >1 acute medication to treat their attacks.
Consistent with previous studies, we demonstrated a much higher incidence of RLS in migraine with aura patients than reported in the general population.3,24 Thirty-eight percent of patients were found to have a large PFO, and 60% had shunts of any type.
The demanding primary end point of complete cessation of migraine headache, which in this study was underpowered, was chosen on the basis of observational studies and ethical considerations that demanded the demonstration of a major clinical effect in a population with severe refractory migraine. A significant effect on this end point and the specified secondary end points was not demonstrated. Exploratory analysis was undertaken when it was evident that 2 statistical outliers accounted for more than one third of the overall migraine headaches experienced. When these 2 patients were removed, the implant arm demonstrated a significant reduction in total migraine headache days, consistent with but not proof of a causal relationship between PFO and migraine with aura. Some patients may benefit from closure, but a potential for short-term deterioration exists in a minority of patients.25 Larger randomized controlled studies that are ongoing will help further define the risk-to-benefit ratio.
Results from the MIST trial did not support the efficacy seen in previous observational reports.6,8–10,18,19,26 A simple placebo response cannot explain the lack of efficacy because patients were not being treated for migraine in the observational studies and therefore had no expectation of efficacy. The discrepancies, however, can be explained in a number of ways. First, in the observational studies, the PFO was closed because it was thought to be responsible for a clinical event, usually stroke or decompression illness,6,8,9,18,19,26 whereas the MIST trial patients were different in that their PFO was not related pathophysiologically to any such clinical event. Indeed the types of patients in the observational studies were specifically excluded from MIST. Second, MIST trial patients were selected because they had particularly severe and refractory migraine, whereas in the observational studies, migraine was incidental to the reason for closure.27 Severe refractory migraine, particularly if associated with chronic frequent headache, depression, or other comorbidities, may prove less amenable to treatment than mild or moderate migraine. Moreover, the continued use of prophylactic migraine medication throughout the trial in both treatment arms (in contrast to most pharmacological studies) may have limited the impact of PFO closure. This patient population typically is excluded from pharmacological migraine trials because they have been shown to be resistant to other drug therapies.
Third, the primary study end point of migraine cessation may have been unrealistic and less clinically relevant than reduction in migraine frequency. Even the best-designed studies of preventive medications show a responder rate (reduction of migraine frequency of ≥50%) of only ≈50%.28 The most commonly used primary end point in such studies is the change in mean monthly migraine frequency, with the responder rate used as a key secondary end point.11,28 In addition, in light of the observed effect size, the secondary end points were underpowered in the MIST trial.
Finally, a number of additional methodological issues may have influenced the results. We chose to analyze the benefit of PFO closure from 3 to 6 months after device implant. The effect of PFO closure during this relatively early analysis phase may have been confounded by a hangover effect of clopidogrel,29 incomplete closure of the defect, concomitant pulmonary shunts, and a possible early transient adverse effect of device implant.25,30 Therefore, a longer analysis phase might have demonstrated additional benefit accrued over time. Residual shunts were assessed by the investigators using contrast transthoracic echocardiography at 6 months. Closure rates were consistent with those previously reported for the STARFlex device.31 However, it is likely that more residual shunts persisted earlier during the analysis phase, and atrial or pulmonary shunts below the detection threshold of this technique32 might have had an impact on the treatment effect in this population.
In 5 patients, the PFO was not crossed. The screening echocardiograms of the patients in whom a PFO was not found were reviewed again, and the conclusions were consistent with the original assessment. The choice of transthoracic echocardiography as a screening method was based on logistical and ethical imperatives, and we believe it has been shown to have acceptable sensitivity and specificity. However, differentiation of the degree and site of shunt may be difficult,33 and additional sources of shunting such as pulmonary atriovenous malformation may be overrepresented in a migraine population. Furthermore, a number of the investigators reported greater difficulty in finding and crossing the PFOs in our study population compared with previous patients with stroke and decompression illness. This might reflect smaller or more serpiginous defects and may have contributed to the adverse events in the study.
It should be noted that the side effects in this trial were transient. Although it is true that discontinuing prophylactic drugs can eliminate side effects, severe persistent side effects are known. The safety profile in this study was consistent with previous reports34 and the known STARFlex safety profile.31 To date, >25 000 PFOs have been closed in clinical practice through 4 generations of technology (NMT Medical Inc, data on file).
The many lessons learned during the conduct and final analysis of this study are crucial to the design of future research. All studies currently approved by the Food and Drug Administration have different study designs with improvements based on lessons from MIST. MIST III, designed to openly follow up patients in the MIST trial, is ongoing, and larger randomized controlled trials with longer-term follow-up are currently underway. Modifications of the patient selection criteria, the primary end point to assess a responder rate, and duration of follow-up, as well as beginning assessments once the implant is fully healed, are some of the necessary changes for new studies.
This trial has confirmed the high prevalence of RLS in migraine with aura patients. Although no significant effect was found for the primary or secondary end points, the exploratory analysis supports further investigation. MIST emphasizes the critical importance of blinding in the evaluation of novel interventions and illustrates that blinding can be achieved even in complex trials. The robust design of this study has served as the model for other larger trials that are currently underway in the United States and Europe.
Professor Horst Sievert, cardiologist chairman of DSAEMB; Professor Eric Eeckhout, cardiologist member of DSAEMB; Professor Len Doyal, medical ethicist member of DSAEMB; Dr Ralph Kern, neurologist member of DSAEMB; Dr Francis Baudet, pain specialist member of DSAEMB; Roy Taylor, biostatistician member of DSAEMB; Dr Luc Missault, cardiologist medical monitor and member of DSAEMB; Geoff Fournie, NMT Medical Inc, member of the steering committee; and Gill Glennon, NMT Medical Inc, member of the steering committee.
Source of Funding
This study was funded by NMT Medical Inc (Boston, Mass).
All study sites received research grants. Drs Hildick-Smith and Mullen have ownership interests in NMT Medical Inc. Dr Mullen has received teaching honoraria and has acted as a consultant to NMT. The remaining authors report no disclosures.
Lipton RB, Liberman JN, Kolodner KB, Bigal ME, Dowson A, Stewart WF. Migraine headache disability and health-related quality-of-life: a population-based case-control study from England. Cephalalgia. 2003; 23: 441–450.
Anzola GP, Magoni M, Guindani M, Rozzini L, Dalla Volta G. Potential source of cerebral embolism in migraine with aura: a transcranial Doppler study. Neurology. 1999; 52: 1622–1625.
Anzola GP, Frisoni GB, Morandi E, Casilli F, Onorato E. Shunt-associated migraine responds favorably to atrial septal repair: a case-control study. Stroke. 2006; 37: 430–434.
Olesen JS, Stephen D. Key issues in the methodology of prophylactic migraine trials. In: Tfelt-Hansen P, ed. Preventive Pharmacotherapy of Headache Disorders. New York, NY: Oxford University Press; 2004.
Headache Classification Subcommittee of the International Headache Society. The International Classification of Headache Disorders: 2nd edition. Cephalalgia. 2004; 24 (suppl 1): 9–160.
Daniels C, Weytjens C, Cosyns B, Schoors D, De Sutter J, Paelinck B, Muyldermans L, Van Camp G. Second harmonic transthoracic echocardiography: the new reference screening method for the detection of patent foramen ovale. Eur J Echocardiogr. 2004; 5: 449–452.
Health Assessment Lab. SF-36v2 Health Survey. Waltham, Mass: Medical Outcomes Trust and Quantimetric Inc; 1992–2002.
Stewart WF, Lipton RB, Dowson AJ, Sawyer J. Development and testing of the Migraine Disability Assessment (MIDAS) Questionnaire to assess headache-related disability. Neurology. 2001; 56 (suppl 1): S20–S28.
Anscombe FJ. Outliers. International Encyclopedia Soc Behav Sci. 1968; 15: 178–182.
Rigatelli G, Cardaioli P, Braggion G, Giordan M, Fabio D, Aggio S, Roncon L, Chinaglia M. Resolution of migraine by transcatheter patent foramen ovale closure with Premere Occlusion System in a preliminary series of patients with previous cerebral ischemia. Catheter Cardiovasc Interv. 2007; 70: 429–433.
Alecrim-Andrade J, Maciel-Junior JA, Cladellas XC, Correa-Filho HR, Machado HC. Acupuncture in migraine prophylaxis: a randomized sham-controlled trial. Cephalalgia. 2006; 26: 520–529.
Anzola GP, Morandi E, Casilli F, Onorato E. Different degrees of right-to-left shunting predict migraine and stroke: data from 420 patients. Neurology. 2006; 66: 765–767.
Wilmshurst PT, Nightingale S, Walsh KP, Morrison WL. Clopidogrel reduces migraine with aura after transcatheter closure of persistent foramen ovale and atrial septal defects. Heart. 2005; 91: 1173–1175.
Interventional Procedure Consultation Document: Percutaneous Closure of Patent Foramen Ovale for the Prevention of Cerebral Embolic Stroke. London, UK: National Institute for Clinical Excellence; 2005.
The Migraine Intervention With STARFlex Technology (MIST) trial was the first randomized controlled clinical trial to evaluate closure of a patent foramen ovale to treat refractory migraine headaches. All other reports to date of migraine improvement after patent foramen ovale closure were on patients with comorbid conditions such as stroke, transient ischemic attack, or decompression illness. The unique study design of the trial demonstrated that a double-blind sham-controlled study was both feasible and ethically justifiable in this condition. Although the study failed to achieve its primary end point of complete cure of recurrent migraine headaches, the modest treatment effect demonstrated in this trial may have been mitigated by a number of confounding factors. The length of follow-up, the assessment period, or the impact of study medications in both arms may have affected the results. Longer-term follow-up of the current study group (including the crossover from the sham arm of the study) and future trials should shed light on the efficacy and risk-to-benefit ratio of patent foramen ovale closure for migraine.
Continuing medical education (CME) credit is available for this article. Go to http://cme.ahajournals.org to take the quiz.
Dr Dowson is the study’s chief investigator.
Drs Dowson, Mullen, and Peatfield are the responsible authors.
Clinical trial registration information—URL: http://www.controlled-trials.com. Registration number: ISRCTN 45687883.
The online-only Data Supplement is available for this article at http://circ.ahajournals.org/cgi/content/full/CIRCULATIONAHA.107.727271/DC1.