Throughout the growth of the child, and continuing into adult life, systemic arteries become less elastic/stiffer and systolic blood pressure rises, the result of increasing systemic vascular resistance [9]. With the higher afterload on the left ventricle (LV), there is myocardial hypertrophy, increased wall thickness, and a reduction in the ease of filling the LV chamber (decreased compliance). As a result, LA pressure will rise. In the geriatric population, there may be loss of myocytes with additional hypertrophy and fibrosis, as well as impairment of myocardial relaxation with changes in the calcium-ATPase pump [10]. All of these further the elevation of LA pressure from its baseline as patients age. In those patients in whom the flap of the PFO has never fused, the PFO will open less frequently as LA pressure rises relative to RA pressure, particularly under normal physiologic conditions. Landeta et al. [11] demonstrated that elevated LVEDP and a dilated LA (as a marker for higher LA pressure) were both negatively associated with the echocardiographic identification of a PFO.

A number of other investigators have confirmed a reduced prevalence of PFO with advancing patient age. Yahia et al. [12], in a retrospective transesophageal echo study, found that patients <60 years of age were identified as having a PFO 33 % of the time, while only 24 % of older patients had a PFO. Fisher et al. [13], in an earlier generation of transesophageal echo technology (presumably less sensitive), also found a significant difference in the prevalence of PFO, with patients 40–49 years of age having a PFO 13 % of the time, while 6 % of patients 70–79 years had a PFO (p = 0.03). These age-related differences could all be related to changes in LV compliance and LA pressure alone, with failure of the PFO to open under the physiologic conditions of a sedated TEE study. These findings do not mean that there is septal fusion and sealing of the PFO that occurs later in life.

However, Hagan et al. [14], in an autopsy series of nearly 1,000 normal hearts, found that in patients ≤30 years of age, 34 % had a PFO. In patients between 31 and 80, the number was reduced to ~25 %, and in patients >80 years old, a PFO was identified in only 20 %. Unlike the echocardiographic series, these findings reflect actual fusion of the PFO flap, not just pressure driven closure of the PFO.

It is unclear why there would be fewer PFOs in older patients. There have been no case reports of late spontaneous PFO closure in adults. Nor is there evidence of a biologic mechanism to explain spontaneous fusion of the PFO outside of the newborn period. A more “Darwinian” explanation is that the absence of a PFO favors longevity based on the mortality associated with potential paradoxical thromboembolism/stroke. There is compelling evidence that the risk of stroke, in older populations with PFO, is increased compared to similarly aged patients without PFO.

Homma et al. [15] retrospectively analyzed patients from the PICCS trial, comparing patients with and without PFO who had presented with a cryptogenic stroke. In patients >65 years of age, the presence of PFO “significantly increased the risk of adverse events, [compared with] the younger cohorts (<55 years, 55–64 years).” While the major adverse events tallied included cardiovascular death, stroke and TIA, the increased rate of stroke in the older patients accounted for all of the difference between the age groups. In patients with recurrent stroke, the adjusted hazard ratio of having a PFO was 4.21 (95 % CL = 1.36–13.02, p = 0.01).

Handke et al. [16] prospectively assessed 503 patients presenting with stroke, of which 227 were deemed cryptogenic. 131 patients were less than 55 years of age, 372 were ≥55. The prevalence of PFO was greater in the cryptogenic stroke population for both younger (43.9 % vs. 14.3 %; odds ratio, 4.70; 95 % confidence interval [CI], 1.89–11.68; P < 0.001) and for older patients (28.3 % vs. 11.9 %; odds ratio, 2.92; 95 % CI, 1.70–5.01; P < 0.001). Multivariate analysis adjusting for age, plaque thickness, presence/absence of coronary artery disease and hypertension showed that the presence of a PFO was independently associated with cryptogenic stroke in both groups (younger cohort: odds ratio, 3.70; 95 % CI, 1.42–9.65; P = 0.008; older cohort: odds ratio, 3.00; 95 % CI, 1.73–5.23; P < 0.001).

Kent et al. [17], in their RoPE metanalysis of stroke-PFO patients, showed that while recurrent cryptogenic strokes are more far more likely to be “PFO-related” in younger healthier patients, the rate of recurrence may be up to ten times as high (20 % vs. 2 % over 2 years) in older PFO patients who have additional traditional stroke risk factors (including hypertension, diabetes mellitus, and smoking). While some of that difference is certainly attributable to the non-PFO risks found in these patients, it does not rule out the possibility that the risk of PFO-related stroke remains significant, as Handke’s data suggested, or even increases with advancing age. It also does not specifically address the issue of cryptogenic stroke in an older PFO population with no other definable risk factors.

It may not simply be the PFO that predisposes any individual to paradoxical embolization/stroke and increased risk of mortality. Rather, normal or abnormal age-related changes in the presence of a PFO may be responsible for higher risk.

Any condition, which would increase the clot burden returning to the RA, makes it more likely that paradoxical embolization would occur in the presence of a PFO. By far the most important of these age-related changes is the increased risk of venous clot formation in the lower extremities and in the pelvis. Reduced activity levels and a higher prevalence of venous valve disease, substantially increase the incidence of lower extremity venous thrombosis in all aging patients [18]. Patients may acquire thrombophilic/hypercoagulable conditions with aging. For example, cancer patients/survivors may develop autoimmune antibodies (anti-phospholipid syndrome) that may promote venous clot formation [19]. The need for estrogen-based therapies (i.e. post-menopause hormone replacement therapy) also increases the venous thrombus risk [20]. Surgical interventions, more common to the older population may also substantially increase risk of venous thromboembolic disease. Joint replacement surgery, for example, is highly associated with lower extremity venous thrombus risk, both from the surgery itself and from the recovery period following surgery [21], while the presence of a transvenous pacemaker/defibrillator has recently been shown to markedly increase the risk of stroke in the presence of a PFO [22].

Any condition which increases streaming of IVC flow toward the fossa ovale could be associated with a higher rate of paradoxical embolization. In a rare clinical entity, “platypnea-orthodeoxia syndrome” [23], right hemi-diaphragmatic elevation, often associated with lung resection, changes the angle of inflow from the IVC such that desaturated blood preferentially streams across the PFO to the left heart. These patients can become profoundly hypoxemic, most often associated with an upright position. We have noted that an unusually high percentage of stroke/PFO patients, also have elevated right hemi-diaphragm (unpublished observation). In these patients, the volume of right to left shunting may not be sufficient to manifest hypoxemia, perhaps because the PFO is smaller, or less mobile. However, the preferential streaming of lower extremity venous return into the fossa ovale may predispose these patients to a higher stroke risk. With aging, diaphragmatic elevation becomes more common. A number of etiologies can lead to phrenic nerve dysfunction including cervical spine injuries/degeneration, infectious diseases such as herpes zoster, Lyme disease, and West Nile Virus, diabetes mellitus, surgical injury, and tumor compression. Eventration of the diaphragm, a congenital thinning of the muscle, is most common in the antero-medial portion of the right hemi-diaphragm. Any increase in intra-abdominal pressure, such as that associated with obesity can cause functional diaphragmatic elevation [24].

Finally, any elevation of RA pressure, relative to LA pressure, will favor right to left flow across the PFO.

A simple Valsalva inhibits systemic venous return to the RA due to raised intrathoracic pressures. Upon Valsalva release, venous return to the RA is augmented with an acute rise in RA volume/pressure [25] and opening of the PFO flap. Valsalva, related to vocational or recreational activities are more frequent in adults than in children, while constipation, leading to strain while moving the bowels, is far more prevalent in older adults secondary to issues such as diabetes mellitus, hypothyroidism, diverticulosis, depression and the use of opiates.

A number of potentially age-related cardiac conditions, including pulmonary hypertension and right ventricular/right coronary artery infarction, may reduce RV output, resulting in higher RV end-systolic volumes, reduced filling and elevation of RA pressure. The severity of tricuspid valve disease may progress with aging and also impact directly on the mechanics of the RA. Worsening tricuspid regurgitation will cause both volume and pressure overload of the RA, while tricuspid stenosis (i.e. endocarditis, carcinoid), creates a pure pressure overload of the RA. Any of these changes can increase the risk of paradoxical thromboembolism across a pre-existing PFO.