Editors' note: Migraine with visual aura is a risk factor for incident atrial fibrillation: A cohort study

In “Migraine with visual aura is a risk factor for incident atrial fibrillation: A cohort study,” Sen et al. followed 11,939 patients with headache and no diagnosis of atrial fibrillation for 20 years and found that after adjusting for confounders, migraine with visual aura was associated with an increased risk of atrial fibrillation. They postulated that autonomic dysfunction may be the underlying cause of both atrial fibrillation and migraine and questioned whether migraine with aura is the result of cardioembolic stroke secondary to atrial fibrillation. Gupta challenged the value of Sen et al.'s findings and commented that (1) migraine with aura and migraine without aura have the same amount of autonomic dysfunction (although no source was provided to quantify the amount of autonomic dysfunction in these 2 entities), and (2) it would be nearly impossible for thromboembolic events due to atrial fibrillation to serially occur in the same cerebrovascular territory leading to migraine with aura. In response, Sen (1) replied that migraine with aura and migraine without aura are commonly considered to be distinct entities and pathophysiologic variants and cited a document published by the International Headache Society and (2) cited a review article that concluded that migraine with aura tends to produce more significant autonomic impairment than migraine without aura. In addition, Sen reinforced that there is a relationship between both (1) migraine with aura and atrial fibrillation (as shown in the present study) and (2) ischemic stroke and migraine with aura (as shown in a previous study). However, it remains unclear whether autonomic dysfunction is responsible for, or merely related to, migraine. Last, Hsieh noted that the x-axis of the Kaplan-Meier curves showing 20-year outcome of incident atrial fibrillation in figure 1 should be labeled “Time to atrial fibrillation,” not “Time to stroke,” and that the log-rank p value of 0.0048 shown on the figure is different from that which is noted in the text (p = 0.0002). Sen replied that Hsieh is correct that the x-axis label should be changed, but said that the p value in the figure is correct (and did not clarify why it is different from the p value in the text).