Fig. 29.1
Illustration of thrombus-in-transit on echocardiography. (a, b) Transesophageal images documenting a large thrombus straddling the patent foramen ovale with extension into both atria. (c) This thrombus was surgically resected from the atria in addition to thrombi from the pulmonary arteries (These images are reprinted with kind permission from Springer Science and Business Media: Bugra et al. [1])
Table 29.1
Clinical associations implicated with patent foramen ovale
Cryptogenic stroke |
Paradoxical embolism |
Migraine |
Decompression illness |
Platypnea orthodeoxia syndrome |
Obstructive sleep apnea |
Hypercoaguable states of the venous system |
Genetic conditions, indwelling pacemakers, varicose veins |
Transient global ischemia |
Deoxygenation from right-to-left shunting |
The awareness of PFO physiology coupled with real-world experience with PFO device closure for clinical conditions such as cryptogenic stroke has led to additional questions. Are there high-risk populations for paradoxical embolism in whom PFO closure may be an appropriate primary or secondary prevention strategy? What are other clinical conditions worthy of investigation? And is there something inherent in the PFO anatomy or physiology that makes patients “arrhythmogenic?”
In the scope of this chapter, we will look at several focus points promising for future PFO research. We will examine data in regards to paradoxical embolism in high-risk populations including hypercoaguable conditions and in patients with indwelling pacemakers. We will review the hypothesis and state of evidence for PFO closure with regards to obstructive sleep apnea (OSA) and what direction future investigation should take. Finally, we will look at what relationship if any exists between the PFO as an independent risk factor for atrial arrhythmias.
Paradoxical Embolism in Hypercoaguable Conditions
In the literature, the contribution of PFO to cryptogenic stroke has received the lion’s share of spotlight. The hypothesis of paradoxical embolism crossing a PFO has been theorized as far back as the late nineteenth century with Drs. Zahn and Cohnheim’s post-mortem discoveries of thrombus trapped across the atrial septum [4–6]. In 1971, Banas et al. published how valsalva can induce right to left shunting in atrial communications of all sizes compared to contrast alone [7]. Multiple modern day publications have documented “thrombus-in-transit” prior to embolic events (Fig. 29.1) [1, 8].
For the past several decades, percutaneous closure of PFO has been performed worldwide both on- and off-label for a number of indications, though most commonly for cryptogenic stroke. The evidence supporting this treatment paradigm has stemmed largely from observational studies and case series; these studies have found device closure to be a safe procedure with low complication rates and suggested a reduction in future embolic events compared to medical therapy alone [6, 9–15]. Only in recent years has data from three large randomized control trials (RCTs) comparing device closure to medical therapy in cryptogenic stroke patients been available [16–18].
The concept of paradoxical embolism as a cause of cryptogenic stroke relies on the formation of systemic venous thrombosis. While only 5–15 % of published PFO cryptogenic stroke trials have documented deep venous thrombosis or pulmonary embolism, most emboli are believed to arise from the intra-abdominal systemic venous system (e.g. pelvic veins) [4, 19]. These venous beds are rarely evaluated as part of a routine stroke evaluation. In other cases, there may be no evidence of residual venous thrombosis as the culprit thrombus has already embolized.
It stands to reason that patients with hypercoaguable disorders that increase the risk for forming venous thrombosis represent a high-risk population for cryptogenic stroke. Polymorphisms of coagulation proteins are common prothrombotic abnormalities that predispose individuals to venous thrombosis (Table 29.2). The most common of these genetic hypercoaguable conditions, Factor V Leiden (G1691A) mutation and Prothrombin (G20210A) mutation have a combined incidence of 2–15 % in the population and carry a three to eightfold increased risk of venous thromboembolism [28].
Table 29.2
Hypercoaguable conditions that predispose towards venous thrombosis
Prevalence in patients with venous thromboembolism (%) | |
|---|---|
Inherited hypercoaguable conditions | |
Factor V Leiden (G1691A) mutation (activated protein C) | |
Homozygous 3 % [21] | |
Prothrombin Factor II (G20210A) mutation | |
Antithrombin III deficiency | 0.5 % |
Methylenetetrahydrofolate (MTHFR C677T) mutation | Homozygous 9 % [25] |
Protein C deficiency | 3 % [26] |
Protein S deficiency | 7 % [26] |
Dysfibrogenemia | 0.8 % [27] |
Acquired hypercoaguable conditions | |
Antiphospholipid syndrome | 4 % [26] |
Acquired hyperhomocysteinemia | Unknown |
Select studies have found hypercoaguable disorders to be associated with increased risk of ischemic stroke, especially in younger patients. One study from Spain examined 100 sequential stroke patients <55 years old [29]. They found 46 % of these patients to have a hypercoaguable state with the most common abnormalities being acquired hyperhomocysteinemia, protein C & S deficiency, factor V Leiden mutation, and methylenetetrahydrofolate (MTHFR) mutation. There was no statistically significant difference in other traditional vascular risk factors noted between the hypercoaguable and non-hypercoaguable patients. A meta-analysis including 56 observational studies and 54,547 patients found factor V Leiden variant, prothrombin mutation, and MTHFR mutation to be associated with increased risk for ischemic stroke and myocardial infarction, especially in women and in patients <55 years [30].
The literature in regards to hypercoaguable states in cryptogenic stroke patients with PFO is limited. Several case series have linked specific hypercoaguable disorders with PFO and stroke [31, 32]. One case control study of 125 consecutive stroke patients (mean age 35 years), found there to be a higher incidence of prothrombin mutation and Factor V Leiden mutation (19 % vs. 3 %) in stroke patients with PFO versus those without [33]. This association of these particular hypercoaguable conditions has been suggested in several studies in varying degrees [34, 35]. A meta-analysis that combined data from 856 PFO-associated stroke cases and 1,001 control subjects found prothrombin mutation to be linked with PFO-associated stroke versus non-PFO-associated stroke (OR 2.3, 95 % CI 1.2–4.4) [36].
Antiphospholipid syndrome (APLS) is somewhat unique amongst hypercoaguable conditions – it is an acquired autoimmune condition that is associated with both venous and arterial thrombosis [37]. After diagnosis, patients are typically maintained on systemic anticoagulation. Several retrospective studies have found APLS antibodies to be common in patients with PFO and a history of systemic thromboembolism [38]. One analysis found the prevalence of PFO with ASA to be significantly higher in an APLS population with stroke compared to a control group of cryptogenic stroke (67 % vs. 20 %, p = 0.015) [39]. However, a secondary prospective analysis of the Patent Foramen Ovalve in Cryptogenic Stroke Study (PICSS) failed to find an increased risk of stroke in APLS patients with PFO compared to those without PFO.
The increased association of hypercoaguable disorders in PFO-related strokes is intriguing but is not decisive in regards to etiology. Patients with diagnosed hypercoaguable states are often excluded from trials and studies examining PFO closure. Part of the rationale has been that hypercoaguable disorders often necessitate anticoagulation, whether or not a PFO is present. There are numerous challenges with anticoagulation including sub-therapeutic levels, poor medical compliance, and comorbid conditions that necessitate anticoagulation interruption. All of these scenarios can leave these patients at risk for embolic events including paradoxical embolism. In patients with thrombophilia’s, PFO closure can be performed safely with equivalent efficacy in comparison to other PFO cases receiving closure for cryptogenic stroke [40].
Summary
Based on available data, there is a role for transcatheter closure of PFO as secondary prevention strategy for hypercoaguable conditions, especially if already on anticoagulant therapy during the index event. The biologic plausibility of paradoxical embolism is based on over a century of real world cases/studies featuring cryptogenic stroke with evidence of venous thrombosis, pulmonary embolism, following intravenous injections, and thrombus-in-transit. Specific hypercoaguable conditions for venous thrombosis have been associated with higher incidence of stroke. The combined weight of research evidence including the observational data, 3 RCTs and published meta-analyses looking at cryptogenic stroke suggests that PFO closure can be safe and beneficial in an appropriately selected population. In patients with hypercoaguable conditions who experience unexplained systemic embolization or cryptogenic stroke, we feel there is already enough evidence and rationale to consider PFO closure.
Further investigation is required before using PFO closure as a primary prevention strategy in patients with venous hypercoaguable conditions (who have a PFO but no evidence of paradoxical embolism). Individuals with genetically confirmed hypercoaguable conditions (both homozygous and heterozygous) are not routinely treated with systemic anticoagulation in the absence of a thromboembolic event or in the presence of high-risk situations (e.g. post-surgical). In the presence of a venous thromboembolic event, anticoagulation remains the mainstay of therapy. However, given the real-world problems with anticoagulation treatment failure, interruption, or non-compliance, there may be a role for device closure on a case-by-case basis. To see a treatment effect in this population, a trial with a large number of patients and long-term follow-up is needed. This represents a subpopulation that is worthy of further trials despite the logistical difficulty.
Chronic Pacemakers or Defibrillators
An increasing number of cardiac patients receive intracardiac leads for pacemakers, defibrillators, or resynchronization devices [41]. While transvenous implantation is generally safe and easy, the intravascular portion represents a foreign body with thrombogenic potential [42]. Venous thrombosis of pacemaker leads is associated with 0–6 % risk of symptomatic upper extremity DVT or SVC syndrome and 0–5 % risk of symptomatic pulmonary embolism [43–46]. Thromboembolic event rates after intracardiac lead implantation may also be under-diagnosed. One study performed V/Q scans in patients before and after pacemaker placement, finding a 15 % incidence of asymptomatic pulmonary embolism [46].
The incidence of thrombus and debris being found on these devices is certainly higher than the incidence of clinical pulmonary embolism (Fig. 29.2). An autopsy study of 79 patients with transvenous pacemakers found high rates of thrombi on ventricular and atrial leads in 33 and 48 % respectively [48]. One study used ICE to assess for lead thrombus in patients about to undergo ablation of atrial arrhythmias. They document mobile thrombi to be present on the intracardiac leads in 30 % of patients. In nearly all these cases, thrombus was not appreciated by transthoracic echocardiography [43].
Fig. 29.2
Echocardiographic illustration showing thrombus lining the pacemaker wire. (a, b) Transesophageal echocardiographic images document thrombus adherent to a transvenous pacemaker wire in a patient with recurrent pulmonary emboli (These images are reprinted with permission from Arslan et al. [47] with permission from BMJ Publishing Group Ltd)
Given thromboembolic concerns, pacemaker leads are avoided in patients with intracardiac shunts [49]. Several case series have now been published describing patients with cardioembolic stroke shortly after pacemaker implantation [50]. Khairy et al. examined a 64 patient cohort with congenital heart disease including intracardiac shunts (PFO was excluded). In multivariate analysis, transvenous leads were an independent predictor of systemic thromboembolism by >2 fold (HR 2, 6, p = 0.02) [51]. In this study population, there was no risk reduction in systemic thromboembolism noted between aspirin or warfarin therapy. A study from the Mayo Clinic retrospectively examined 6,075 patients receiving pacemaker leads over a decade – from which 364 had echo-documented PFOs. At a median 5-year follow-up, 8.2 % of patients with PFO had evidence of cardioembolic stroke or TIA versus 2 % of non-PFO patients (HR 3.5, 95 % CI 2.3–5.3) [52]. The Mayo study suggests >3 fold risk of stroke or TIA following device implantation with a PFO versus without.
Summary
With regards to pacemaker leads, patients with cryptogenic stroke should be examined on a case-by-case basis. Creating a nidus for thrombus in patients with intracardiac shunts including PFOs is of concern based on the published data to date. At this juncture, it would be reasonable to study this principle in a prospective fashion. It is possible that percutaneous PFO closure may lead to reduction in systemic embolic events by preventing paradoxical embolism. Such a study would require large numbers, but the potential implication may be to avoid endovascular leads in patients with significant PFOs or alternatively consider anticoagulation or selective PFO closure.
Obstructive Sleep Apnea
OSA is defined by the periodic reduction or cessation of respiration during sleep secondary to intermittent upper airway obstruction. OSA is a chronic condition with a high population prevalence between 5 and 15 %, affecting men more often than women [53, 54]. It is the most common medical disorder leading to daytime drowsiness and is associated with a host of other debilitating symptoms including memory impairment, headaches, difficulty with concentration, loss of libido, sensation of choking/gasping, and depression. OSA has been implicated in the development of cardiovascular disease including hypertension [55], coronary and vascular events [56], atrial fibrillation [57], congestive heart failure [58], ventricular arrhythmias [59], and pulmonary hypertension [60]. It has also been associated with an increased risk ischemic CVA events [61]. OSA contributes to considerable functional impairment and is associated with an overall increased mortality [56, 61].
High Rates of PFO Detection in OSA Patients
The idea that PFO may play a role in the pathophysiology of OSA has stemmed largely from observational data over the past several decades. Several studies have found there to be an increased prevalence of PFO in patients with diagnosed OSA. In 1998, Shanoudy et al. published that 69 % of 48 consecutive OSA patients were found to have a detectable PFO on transesophageal echocardiography (TEE) with Valsalva manoeuvre compared to only 17 % of 24 controls (p < 0.0001). Beelke et al. examined 78 patients affected by moderate-to-severe OSA and 89 controls. Using transcranial Doppler (TCD) detection, they document a smaller association of PFO of 27 % in OSA patients versus 15 % in the control group (p < 0.05). A third and larger case control study by Lau et al. from 2013 confirmed this association examining 102 OSA patients compared to 50 controls. Unlike the previous two studies, this study group excluded OSA in their control arm by a negative sleep study. With an average age of 50 years, they report 47 % PFO prevalence in OSA patients compared to 26 % in controls (p-0.014). A fourth study with broader inclusion criteria also examined 100 OSA patients compared to 50 control subjects and found a more modest PFO prevalence of 43 % with OSA and 30 % in controls (p = 0.16) [62].
While there is considerable variability of PFO detection rates between the studies – all four suggest higher PFO prevalence in OSA compared to controls. These inter-study differences may be explained in part by selection bias. All the studies were performed in different countries, and Beelke et al. excluded patients with prior stroke from their study. Differences in PFO detection may play a role – as other trials have suggested TEE with Valsalva to have higher sensitivity in PFO detection than transthoracic echocardiography (TTE) or TCD [63, 64]. It is also possible that increased right atrial pressures in OSA lead to higher PFO detection from increased shunting, than in non-OSA patients [65]. Statistically significant age differences between OSA and control groups may exaggerate the differential. The study by Lau et al. features the largest study population and excluded all patients with history of stroke, migraine, paradoxical embolism, or decompression sickness. Our conclusion from these studies is that individuals with OSA have a 2–4× increased prevalence of PFO than those without OSA.
The PFO-OSA Hypothesis
From a physiology perspective, OSA may duplicate the hemodynamics most likely to induce right-to-left shunting across a PFO. The intrinsic anatomy of PFO allows for variable severity and direction of shunting depending on the relative intracardiac and intrathoracic pressures [66]. Thus physiologic changes can create shunting even in typically “silent PFOs” with little resting shunt. Valsalva manoeuvre involves forceful expiration against a closed mouth/nose, leading to an increase in the intrathoracic pressure, a secondary increase in right atrial pressure, and a decrease in venous return. The inspiratory equivalent of the Valsalva, the Mũller manoeuvre involves inspiration with a closed mouth/nose leading to a sharp fall in intrathoracic pressures and increased venous return. The sudden changes in intrathoracic pressures with Valsalva/Muller can lead to an increased pressure differential between the right and left atrium leading to evidence of shunting across a PFO on corresponding imaging (Fig. 29.3) [67, 68].
Fig. 29.3
Pathophysiology of Obstructive Sleep Apnea (OSA) implicated in Paradoxical Embolism. (a) This schematic shows changes in relative atrial pressures at rest and during periods of elevated right atrial pressure (e.g. apnea spells) that can lead to increases in right-to-left shunting across a patent foramen ovale. RA right atrium, RV right ventricle, LA left atrium, LV left ventricle (b) The intrathroracic pressure swings intrinsic in OSA are hypothesized to worsen right-to-left shunting thereby worsening hypoxemia and the clinical sequelae of OSA (These images are reprinted with permission from Lau et al. [67] with permission from Elsevier)
OSA can recreate this physiology with inspiration/expiration against upper airway occlusion that occurs during apnea episodes. This can lead to the atrial pressure differentials facilitating flow across a PFO (Fig. 29.2). In addition to changes in atrial pressure, apnea episodes are also associated with shifting of the ventricular septum on echo and corresponding finding of pulsus paradoxus [67, 69, 70]. In large PFOs with atrial septal aneurysms (ASAs) and significant resting shunt, apnea episodes will not only increase the shunt severity but also can bring out right to left shunting in an otherwise “silent PFO.” A mechanistic study performed by Belche et al. found PFOs that are clinically “silent” while the patient is awake or during normal tidal breathing sleep, can exhibit right to left shunting with saline injection during frank apnea episodes in OSA [65].
In addition to changes in intra-thoracic pressure during apnea episodes, chronic OSA is associated with pulmonary hypertension in 15–20 % of patients [60, 68]. Pulmonary hypertension can lead to an increase in right ventricular and right atrial pressures thereby increasing right-to-left shunting across a PFO. Pulmonary hypertension is especially noted in OSA patients with nocturnal and daytime oxygen desaturation [71]. While causality has not been established, pulmonary pressures can decrease in OSA patients treated with continuous positive airway pressure [72]. In animal studies, occluding the airways of anesthetized dogs leads to a transient rise in PA pressures. Desaturation intrinsic with apnea episodes can lead to pulmonary vasoconstriction that can further lead to elevation in pulmonary pressures and increased right-to-left shunting [67, 73].
PFO Contributing to Desaturation in OSA
As discussed, there is biologic plausibility of increased right to left shunting across the PFO in the setting of OSA. The clinical implication follows that this increased shunt may lead to further systemic desaturation beyond what occurs from the apnea episodes alone. This deoxygenation coupled with inflammation and oxidative stress is theorized to trigger the inflammatory cascade leading to endothelial and autonomic dysfunction. This further perpetuates the severity of OSA in addition to its host of associated comorbid conditions.
While the increased prevalence of PFO in OSA patients has been established, the evidence in support of desaturation from PFO in OSA is ambiguous. A case control study by Johansson et al. looked at 209 individuals with OSA defining the severity of systemic desaturation using a ratio of oxygen desaturation index (ODI) / apnea-hypopnea index (AHI) [74]. In this study, 15 patients with high proportional desaturation defined as an ODI/AHI ratio ≥0.66 were compared to 15 matched controls with low proportional desaturation, ODI/AHI ratio ≤0.33. Sixty percent of the cases with high proportional desaturation were found on TEE with saline injection to have a large PFO (defined as ≥20 bubbles crossing across the septum) as opposed to 13 % of controls. This provides indirect evidence that patients with large PFOs in OSA are prone to desaturation from shunting. Another observational study found OSA patients with large PFO shunts (by bubble TTE study) to have higher ODI/AHI ratios than those OSA patients without large shunts [62].
Both of the above studies have been criticized for the use of ODI as a surrogate for nocturnal hypoxemia. While this ratio is used in the OSA literature to help quantify the severity of apnea spells, it may be better suited for determining the frequency of desaturation events as opposed to desaturation severity (which may be more relevant in this case). Neither study was able to document differences in nocturnal hypoxemia between cases and controls [62, 74]. Lao et al. performed a regression model to examine if right to left shunt severity (using TCD) impacted directly on the severity of desaturation in their 107 patient series. Using their model, they were not able to find a direct relationship between PFO shunting and nocturnal hypoxemia levels. At this stage, there have not been adequately designed and powered studies published that prove increased desaturation occurring from right to left shunting during OSA apnea episodes.
The Data on PFO Closure for OSA
Over the past decade, both anecdotal evidence and published case reports have described an improvement in OSA symptoms and disease severity after device closure of PFO [75, 76]. One case series examined three patients with severe OSA (as assessed by AHI and ODI values on polysomnogram) in whom PFO closure was performed using the Coherex FlatStent EF PFO Closure System (Coherex Medical, Salt Lake City, USA) [77]. One of the three treated patients had benefit in his OSA symptoms with improvement in daytime somnolence and headaches, lower arterial blood pressure, and no further need for CPAP. Follow-up sleep study showed a reduction in the AHI by ~50 %. The other two patients had experienced no change in symptoms on follow-up – complete PFO closure was achieved in all three patients per follow-up imaging. One case control study describes percutaneous closure of PFOs in 6 of 18 OSA patients with large right to left shunts with Valsalva. Of note at 1 year, 2 of 6 treated patients still had residual large shunts. In these patients they did not find any significant change in AHI, ODI/AHI ratio, 6-min walk test, or CPAP use in follow-up.
Summary
The association and proposed mechanism linking OSA and PFO is intriguing though at this stage further clinical evidence is needed. Future trials should focus on confirming the physiology of increased shunting leading to exacerbation of OSA events. If that is established, then RCTs comparing device closure to standard medical therapy alone would be an appropriate next step. While the evidence remains preliminary, PFO closure might represent a clinical boon in the treatment strategy for select OSA patients.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
