Mas et al. [25] first demonstrated a higher recurrent stroke risk when an atrial septal aneurysm was present in association with a PFO. With a “floppy” septum primum, the right atrial pressures required to open the flap of the PFO would be much less than in a more rigid walled septum. In the former, the earlier depolarization of the right atrium, or the mild increase in systemic venous return with inspiration may be sufficient to generate a right to left shunt, while in the former a Valsalva and release may be required. It follows that when an atrial septal aneurysm is present, the PFO will be open more frequently, and potentially to a larger size for any given degree of right atrial pressure rise.

While this relationship has not always been consistent throughout the literature, the RESPECT Trial [24] also found this to be a statistical predictor of lower recurrent stroke rate with transcatheter closure (OR = 0.19 (95 % Cl 0.04–0.87, p = 0.02). It is important to note that the presence of an atrial septal aneurysm has never been implicated independently, in any series, as a risk for primary stroke.

The more time that a right to left shunt is present, the more opportunity there might be for a thrombus to cross to the left heart. Rigatelli et al. [26] studied the risk of recurrent paradoxical stroke in 320 consecutive patient (mean age 44 ± 10 years.) referred for PFO closure, specifically related to the persistence of the right to left shunt. They defined two populations of patients, those with continuous or “permanent” right to left shunt (N = 180), and those with right to left shunt with Valsalva only (N = 140). The permanent shunt group had an increased frequency of multiple ischemic MRI lesions, prior clinical stroke, previous systemic embolization, and were more often associated with atrial septal aneurysm and a prominent Eustachian Valve. A permanent shunt conferred a higher risk of recurrent stroke (odds ratio 5.9, 95 % confidence interval 2.0–12.0, p < 0.001). On follow-up, no difference existed after closure of the PFO. This finding has also never specifically been assessed as a predictor of primary stroke.

The presence of a prominent Eustachian valve (EV) has been highly associated with both the presence of a PFO as well as an increased risk of paradoxical embolization [27, 28]. The EV is a fetal structure that directs high oxygen umbilical venous blood from the inferior vena cava across the PFO to the left heart. In most patients, EV tissue is resorbed, but in some adult patients, the remnant of the EV can be quite prominent, and continue to direct blood returning from the IVC toward the fossa ovale. In the presence of a significant PFO, the blood returning from the IVC, containing potential thromboembolic material could be preferentially directed to the LA.

Based on the preceding discussion, we could define patients with any of the following findings as a “higher risk” for paradoxical embolization:

Should we consider prophylactic therapy for this population? If so, what therapy is appropriate, blood thinner therapy or closure of the defect? There is virtually no information in the literature concerning the potential benefits of treating asymptomatic PFO patients.

In a unique study, Krasuski et al. [29] retrospectively reviewed surgical outcomes for patients identified with incidental PFO (by TEE imaging) at the time of a scheduled cardiothoracic surgery at the Cleveland Clinic. Over an 11-year period, 2,277 patients were identified with PFO, which had not been previously diagnosed. Six hundred and thirty-nine (28 %) were surgically repaired at the time of the scheduled surgery. Compared with the patients in whom the PFO was left alone, those undergoing closure had a longer cardiopulmonary bypass time (110 vs. 104 min, p = 0.001) and an increased risk of in-hospital stroke (2.8 % vs. 1.2 %, p = 0.04), the latter perhaps related to thrombotic complications at the atrial surgical site.

The three randomized prospective PFO-stroke recurrence trials published since 2012, (RESPECT [24], CLOSURE I [30], PC Trial [31]), have shown no conclusive evidence that there is a significant benefit to transcatheter closure compared with chronic blood thinner therapy for prevention of recurrent stroke. The rate of recurrent stroke in the medical/control arms of these trials was quite low, in what is a historically higher risk population. It follows therefore that if therapy is being considered, for high-risk features, in the “incidental” PFO population, anti-platelet therapy should be the treatment of choice.

For the patients who have none of the “high-risk” features in their PFO profile, we can feel much more confident in our ability to predict a benign course for the PFO. These patients should not require neurologic imaging studies, nor further cardiovascular work-ups. They require no blood thinner therapy. They need no long-term follow-up of the PFO itself.