Letters to the Editor Migraine With Aura and Patent Foramen Ovale: Which is Their Relationship? We read with great interest the study by Domitrz et al 1 evaluating the occurrence of patent foramen ovale (PFO) in patients with migraine with aura (MA) and the relation- ship of PFO with clinical picture of MA (different types of migraine aura and frequency of migraine attacks). The authors found high prevalence of PFO in patients with MA, but no link between the occurrence of PFO and the fre- quency of MA attacks. These data are in agreement with the results reported in one of our previous studies 2 and with a recent study by Tembl-Ferrairo et al. 3 Briefly, we have evaluated 30 con- secutive patients with typical aura and positive for migraine headache using contrast transcranial Doppler evaluation for right-to-left shunt (RLS), and investigated the relation- ship between the extent of the RLS (by means of the number of microbubbles – MBs – detected during trans- cranial Doppler with intravenous injection), and the clinical characteristics of migraine aura (age at first onset of migraine, mean annual frequency of attack, and mean dura- tion of the aura phase). No correlation has been docu- mented between the number of MBs and the above- mentioned clinical parameters. Also, taking into account the data from Domitrz et al, 1 we can imply that PFO does not seem to affect the clinical manifestation of migraine aura and that the occurrence of PFO (and/or its gravity) does not seem to be correlated with the severity of the clinical picture of the MA. What does this mean? The answer may be suggested following some consideration. The absence of any correlation between the occur- rence PFO and the clinical severity of MA might suggest the absence of any pathogenic link between said conditions. As a matter of fact, they might only represent the result of a common genetic substratum that could lead, on the one hand, to a predisposition for migraine and, on the other, to endocardiac alteration with persistence of the foramen ovale. Indeed, a recent report documented that the occur- rence of atrial shunts was consistent with autosomal domi- nant inheritance in some families with MA. 4 However, in one recent study, Tembl et al 5 have reported that patients with migraine who have a large RLS tend to recognize activities increasing the extent of the shunt as triggers of their migraine attacks, thus pro- viding some support for a potential pathogenic correlation (causal relationship), even if, as reported by the same authors, in the study at issue a selection bias cannot be excluded. Furthermore, whether or not the RLS triggers the migraine when a specific “threshold” level is exceeded, or, through chronic shunt, induces a state of central hyper- excitability that is affected by other triggers, is a matter for speculation. 5 What are the therapeutic implications of such observations? Even though it has been suggested that PFO closure may represent a potential therapeutic approach for MA, 6,7 expert opinion underlines the fear that such reports might prompt a flurry of inappropriate PFO closure procedures for the prevention of migraine itself. 8,9 To date, from the first (as yet unpublished) randomized trial to compare PFO closure with a sham procedure for the treatment of MA, there is insufficient evidence that PFO closure has any ben- eficial effect on migraine frequency, severity, or its natural history. 9,10 Speculation on the pathogenesis is outwith our inten- tion, even if, in conclusion, current data from the literature seem to suggest the likelihood of the hypothesis that both conditions could be dominant and share a common genetic background. Until further evidence from con- trolled clinical trials becomes available, PFO closure should not be performed in clinical practice for the prophylaxis of migraine, 8,11 even if the debate on the nature of the relationship between RLS and MA remains open. Nicola Morelli, MD Antonio Tartaglione, MD, Prof. Sant’Andrea Hospital – Neurology Division, La Spezia, Italy Sara Gori, MD, PhD 637