RLS is an abnormal pathway between the venous and arterial circulations, and includes both intracardiac and extracardiac RLS. Intracardiac RLS are usually related to PFO, which has been described as a “back door to the brain” [17, 18]. Hagen et al's autopsy study on 965 normal hearts found that PFO had a prevalence of 27.3% for all age [19]. Therefore, PFO should be regarded as a normal structural variant even without paradoxical embolism or other discomfortable clinical conditions existed [20, 21]. Common detection methods of PFO include c-TTE, c-TCD and Contrast transesophageal echocardiography (c-TEE). Each of them has advantages and limitations in the diagnostics of a patient with stroke, but they should, in principle, be equivalent: they all detect RLS by MBs, either visually (TTE and TEE) or by Doppler shift (TCD). Therefore, provided that the technique and visualization is adequate, and the physiology behind the presence of the shunt (at rest or with the VM) is similar, there should be no significant difference in diagnostic accuracy [22]. C-TEE is regarded as the “gold standard” for the diagnosis of cardiac RLS [23].However, small PFO is hard to detect by TEE, and it may be difficult to perform the VM during TEE, especially in elderly patients with severe neurological deficits [24]. Besides, other less invasive techniques, such as TTE and TCD, have been improved and become preferred method to detect for PFO [12].
In this study, the incidence of permanent and total RLS in migraine was higher than that of controls. In addition, compared with controls, the positive rate of RLS in both MA and MO increased (P < 0.05), this is consistent with Yang et al's study [25]. It suggested that PFO associated RLS may be associated with migraine (MA and MO). However, Yang only used c-TCD to evaluate the positive rate of RLS, which was difficult to determine the anatomic origin of intracranial MES, in other words, it can’t exclude extracardiac RLS, and to some extent it affects the accuracy of the results.12 However, c-TTE can make up for c-TCD’s deficiency and has advantages on the judgment of intracardiac RLS, therefore the results are expected to be more accurate. The mechanism of PFO causing migraine is unknown, possible causes being coughing, straining to defecate, during the VM, and lifting heavy objects which can cause right atrial pressure (RAP) to be increased and exceed the left atrial pressure (LAP), making it easier for RLS to go through a PFO, MES and metabolite products from venous circulation enter the intracranial artery and cause brain stimulation [12]. Nozari et al [26] showed in mice that small particulate or air emboli injected into the carotid artery were able to evoke a cortical spreading depression (CSD) without causing ischemia. This study suggested that abnormal microembolism and ischemia may provide a trigger for migraine [27–29]. In addition, the correlation between PFO and migraine may be related to genetics. A recent report documented that the occurrence of atrial shunt was consistent with autosomal dominant inheritance to some families with aura migraine [30].
In this study, the proportion of large shunt in the MA group and the MO group were increased (especially in MA) compared with controls, suggesting that the correlation between PFO and migraine may be related to large shunt. Larger RLS may increase migraine probability, suggesting a “neuronal threshold” above which migraine is triggered. Jesurum et al’s [31] reported a follow-up study of 67 migraineurs’ migraine symptoms after transcatheter PFO closure, which use migraine relief (> 50% reduction in frequency) as the endpoint, migraineurs with aura were 4.5 times more likely to experience migraine relief than migraineurs without aura. Although some patients have RLS shunt, there was no statistically significant difference in migraine symptom relief between the complete group and the incomplete group (77% vs 83%, P = 0.76). In conclusion, migraine relief may occur despite residual RLS after transcatheter PFO closure, which may suggest a reduction in RLS burden below a neuronal threshold that triggers migraine. Some scholars have reported that migraineurs with RLS were associated with impairment of dynamic cerebral autoregulation (dCA). Guo et al [32] divided 66 migraine patients into the RLS group (n = 30) and the non-RLS group (n = 36). It was found that phase difference (PD) of patients in the RLS group were significantly lower than those in the non-RLS group (P < 0.001), and The PD in the large RLS group was significantly lower than that of the small RLS group (P < 0.01) and non-RLS group (P < 0.001), dCA was impaired in migraineurs with large RLS, and this may represent a potential mechanism linking RLS and migraine. Transcatheter PFO closure has recently become an effective therapy for improving migraine symptoms and reducing ischemic events. It is essential to analyze PFO characteristics and identify high-risk PFO. Compared with controls, the prevalence of large-size PFO (≥ 2.0 mm) increased (P = 0.048), length of PFO, long-tunnel PFO (≥ 10.0 mm) was no difference among groups (respectively, P = 0.199, P = 0.095). It was suggested that PFO may cause migraine related to the large-size PFO. This may be due to an increased risk of paradoxical embolism in large PFO [33].
These discovery of a possible link between migraine and PFO remains controversial. As a rule, embolic events show an unpredictable hemispheric distribution, while migraine pain is typically lateralized, often periodic and predictable, like menstrual migraine. Moreover, patients with PFO combined with migraine experienced a decrease in headache symptoms with age [30]. Whether the relationship between PFO and migraine is causal or symbiotic remains to be studied.
Our study may have some limitations: first, some of the studied subjects were outpatient patients, which may have selection bias. Secondly, the diagnostic method of PFO is c-TEE. Due to invasive examination and difficulty in VM, only some patients have completed c-TEE.