The objectives of this study were to evaluate the incidence, predictive factors, and duration of migraine headache attack (MHA) after transcatheter atrial septal defect (ASD) or patent foramen ovale (PFO) closure. A total of 260 consecutive patients who underwent ASD or PFO closure in our center answered a structured headache questionnaire focused in 3 period times, including (1) at baseline (just before closure), (2) within the 3 months after ASD-PFO closure, and (3) at the last (median 27 months, range 6 to 80 months) follow-up. All questionnaires were evaluated by a neurologist who established the diagnosis of MHA with or without aura, according to International Headache Society criteria. The Amplatzer ASD or PFO device was used in 95% of the patients, and aspirin, for at least 6 months, was the antithrombotic treatment in 91% of the cases. A total of 185 patients (71%) had no history of MHA before ASD-PFO closure, and these constituted the study population (mean age 39 +/- 21 years). MHA occurred in 13 patients (7%) after ASD-PFO closure, with aura in 9 of them. MHA appeared after a median of 10 days (range 0.3 to 30 days) after the procedure and were still present at the last follow-up (23 +/- 17 months) in 9 patients (69%). The median number of MHA within the 3 months after the procedure was 4 per month (interquartile range 1 to 23), and decreased to 1 per month (interquartile range 0.3 to 1) at the latest follow-up (p = 0.03). Compared with the patients who did not develop MHA, patients with MHA after ASD-PFO closure were younger (26 +/- 16 vs 39 +/- 21 years; p = 0.02) and were more likely to have undergone ASD closure (100% vs 58%; p = 0.001). In the multivariate analysis, ASD closure was the only predictor of MHA occurrence after the procedure (odds ratio 7.7; 95% confidence interval 1.5 to 22; p = 0.01). In conclusion, MHA, mostly with aura, occurred in 7% of patients after transcatheter ASD-PFO closure and persisted in most of them after a mean follow-up of 2 years. ASD closure was the only independent predictor of MHA occurrence after the procedure. These results suggest that mechanisms other than device composition are involved in the occurrence of MHA in these cases.