Summary
Patent foramen ovale and obstructive sleep apnoea are frequently encountered in the general population. Owing to their prevalence, they may coexist fortuitously; however, the prevalence of patent foramen ovale seems to be higher in patients with obstructive sleep apnoea. We have reviewed the epidemiological data, pathophysiology, and the diagnostic and therapeutic options for both patent foramen ovale and obstructive sleep apnoea. We focus on the interesting pathophysiological links that could explain a potential association between both pathologies and their implications, especially on the risk of stroke.
Résumé
Le foramen ovale perméable et le syndrome d’apnées obstructives du sommeil sont fréquemment rencontrés dans la population générale. En raison de leur prévalence, ils peuvent coexister fortuitement mais la prévalence du foramen ovale perméable semble néanmoins être plus élevée chez les patients présentant un syndrome d’apnées obstructives du sommeil. Nous passons en revue dans cet article la physiopathologie, le diagnostic et les options thérapeutiques du syndrome d’apnées obstructives du sommeil ainsi que du foramen ovale perméable. Nous mettons l’accent sur les liens physiopathologiques qui peuvent suggérer une potentielle association entre ces deux entités et leurs implications, principalement dans le risque d’accident vasculaire cérébral.
Background
PFO and OSA, both of which are common in the general population, have mostly been studied separately. During the past decade, several studies suggesting an association between PFO and OSA have been published. Our aim is to provide clinicians with a review of all the available data concerning this potential association.
First, we will define PFO and OSA, and summarize the available epidemiological data. The diagnostic criteria and the techniques of detection are discussed. We will then focus on the interesting potential pathophysiological links between PFO and OSA, which should encourage recognition of this association. The specific case of stroke is discussed as a potential dangerous consequence of this association. We conclude with the clinical implications and possible future treatment strategies.
Patent foramen ovale: definition and epidemiology
PFO is an embryological remnant of the foetal circulation caused by incomplete fusion of the septum primum and secundum. This valve-like opening represents the most frequent interatrial communication. The prevalence of PFO is 9.2–24.3% based on TOE , 10–18% on TTE , and 14.6–27.3% on autopsy studies . Based on all available series, the prevalence of PFO in the general population is estimated to be 10–30%. The prevalence of PFO decreases with age, from 34.3% during the first three decades of life to 20.2% during the ninth and tenth decades . However, the size of the foramen tends to increase with age, suggesting that small PFO may seal with time .
In most cases, PFO remains asymptomatic, but it may also have important clinical consequences. No specific symptoms, other than the POS in severe cases, are directly related to PFO, but when right atrial pressure exceeds left atrial pressure (e.g. Valsalva), right-to-left shunting may occur, allowing venoarterial (“paradoxical”) systemic embolization and passage of deoxygenated blood in the left atrium.
Patent foramen ovale and associated diseases
Despite numerous studies, a direct relationship between paradoxical embolization through PFO and stroke remains difficult to prove. A meta-analysis of case-control studies in patients with otherwise unexplained (“cryptogenic”) stroke showed, in 2000, that PFO was more prevalent in such patients, particularly in the young . Another more recent study extended this observation to patients > 55 years of age who suffered stroke (28.3% vs 11.9%; P < 0.001) .
It has not, however, been possible to show that patients who have never had a stroke are at increased risk for a first event in the presence of a PFO and/or an ASA . This may be due to the low relative stroke risk of a PFO compared with other stroke risk factors. Larger cohorts with longer follow-up would therefore be required to show such a correlation . The risk of recurrence after a cryptogenic stroke is also lower than after cardiac embolism (e.g. atrial fibrillation) or after stroke due to large artery atherosclerosis. Moreover, it has been difficult to prove that patients with a PFO and otherwise cryptogenic stroke have an increased recurrence risk . The risk of recurrence after cryptogenic stroke might even be the same as in a patient without PFO. Finally, the presence of an ASA, a large right-to-left shunt, and the coexistence of prothrombotic states seems to increase the association between PFO and cryptogenic stroke, but this has not been uniformly proven .
The potential correlation between migraine and PFO is also intensely debated, and illustrates the difficulty in assessing an association between PFO and another definite pathology with epidemiological studies. The prevalence of PFO in patients who have migraine with aura seemed to be higher (28–48%) than in controls . After encouraging reports in case series indicating a potential benefit on migraine recurrence after PFO closure , the first randomized, prospective, sham-control trial could not confirm these effects .
PFO has also been implicated in decompression sickness. Divers with PFO have a higher load of small ischaemic brain lesions . Also, the prevalence of PFO in divers who have already experienced a major event of decompression illness is higher than in divers who have not; and small PFO seem to present less risk of decompression illness .
Obstructive sleep apnoea: definition and epidemiology
OSA is a common sleep disorder, described as repeated closure of the upper airway during sleep. The obstruction may be due to: a decreased activity of the pharyngeal musculature; pharyngeal fat deposits; and mucosal inflammation leading to occlusion or near occlusion of the upper airway . OSA is usually defined as more than five episodes of apnoea or hypopnoea per hour of sleep. The AHI is usually used to assess the severity of sleep apnoea (mild: 5–15; moderate: 15–30; severe > 30 events/h) . Young et al. studied a middle-aged population and described a prevalence of OSA (AHI ≥ 5) of 9% inwomen and 24% in men . Patients with sleep apnoea syndrome have AHI ≥ 5 and symptoms during the day, principally hypersomnolence due to sleep fragmentation. The prevalence of this syndrome in the study byYoung et al. was 2% in women and 4% in men . It is also associated with an increased incidence of hypertension, cardiovascular disease and traffic accidents . OSA has also been reported in association with nocturnal arrhythmias and stroke . OSA is frequently undiagnosed, leading to a substantial cardiovascular morbidity.
Obstructive sleep apnoea and risk of stroke
Case-control and cross-sectional studies have shown an association between snoring and cardiovascular disease . An association between sleep-disordered breathing and stroke has also been evoked for many years . In 2005, Arzt et al. published cross-sectional and longitudinal analyses of 1475 members of the general population . They found that an AHI ≥ 20 increased the chance of stroke compared to an AHI ≤ 5 (OR 4.33, 95% CI 1.32–14.24; P = 0.02) . In the prospective part of the study ( n = 1189), the adjusted OR for a first-ever stroke in the subsequent 4 years in a patient with AHI ≥ 20 vs ≤ 5 was 3.08 ( P = 0.12) . Although the strong correlation between OSA and cardiovascular disease may in part be due to overlapping risk factors for atherosclerosis , there is evidence that OSA is an independent risk factor considering the potential mechanisms increasing the risk of stroke in OSA patients .
Several hypotheses may explain the increased risk of stroke in OSA patients . Hypoxaemia and decreased cerebral perfusion pressure (due to increased intracranial pressure at the end of the apnoea) are considered to be determinants of cerebral nocturnal ischaemia. Haematological parameters may also play a role, such as enhanced platelet aggregation, increased viscosity, and higher fibrinogen concentration . This increase in fibrinogen may decrease with CPAP . Cardiac arrhythmias are more frequent in OSA patients, particularly atrial fibrillation, which is a risk factor for stroke .
Diagnosis
The prevalence of a disease is dependent of the definition and the technique of detection used . The choice of the diagnostic tool and a strict protocol are crucial to diagnose PFO and OSA.
Diagnosis of patent foramen ovale
PFO is usually diagnosed with TTE or TOE, with injection of agitated saline contrast to create microbubbles. Contrast TOE is considered as the gold standard. The definition of PFO is based on echocardiographical criteria: ≥ 3 bubbles (≥ 1 according to some authors) must be observed in the left atrium within three cardiac cycles after the entry of the bubbles in the right atria. Direct visualization of bubbles passing through the PFO channel may also be considered as proof of PFO.
The sensitivity of microbubble assessment of PFO is increased by provocative manoeuvres such as cough and voluntary Valsalva . At the end of Valsalva strain, a transient pressure elevation inside the right-sided cardiac chambers is observed, increasing the likelihood of a right-to-left interatrial shunting. A correct Valsalva manoeuvre is crucial and the patient must receive appropriate instructions before the test. A leftward displacement of the interatrial septum indicates a higher right than left atrial pressure . Increasing the number of injections, and early rather than late Valsalva, is more effective . Usually, injection of contrast is performed antecubitally, but femoral injection has also been used .
Contrast-enhanced transcranial Doppler is also a valuable tool for PFO diagnosis, allowing noninvasive examination of the middle cerebral artery with good sensitivity and specificity. It is considered as efficient as echocardiographic techniques (evidence level type A, class II) . This technique allows the examination of patients while asleep. However, it does not allow differentiation of the source of shunting (intrapulmonary or intracardiac) or provide any information on the morphology of the interatrial septum (e.g. ASA) and on other intracardiac source of embolization (e.g. thrombus) .
Diagnosis of obstructive sleep apnoea
The gold standard method for the diagnosis of OSA is polysomnography: electroencephalogram, electrocardiogram, electrooculogram, submental and tibial electromyograms, body position detector, nasal pressure recording (or an oronasal thermistor), pulse oximeter, and belts to monitor movements of the chest wall and abdomen. In some European countries, and recently in the USA, limited channel home monitoring (type III) is considered as a suitable alternative when pretest clinical suspicion of OSA is high. This type of monitoring is much cheaper and more convenient (as it can be performed at home), but is less accurate as there is no electroencephalogram.
Coexistence of patent foramen ovale and obstructive sleep apnoea
Owing to their high prevalence in the general population, it is evident that PFO and OSA may coexist fortuitously in many patients. A strict methodology is thus required to assess a potential relationship between PFO and OSA to avoid bias. The controversies observed in the studies concerning stroke and migraines are an important illustration of this problem. Two studies have specifically examined the prevalence of PFO in OSA patients. Both concluded that PFO is more frequent in patients with OSA than in controls, but with important differences in their results. The first study, conducted by Shanoudy et al., used TOE in 72 awake male patients . They detected 33 PFO in 48 patients with OSA and four PFO in 24 controls (68.8% vs 16.7%; P < 0.001) . In this study, mean AHI was 33.9; and mean pulmonary artery systolic pressure was higher in patients with OSA than in controls. This study also revealed that one-third of patients with OSA and PFO had a significant decrease in oxygen arterial saturation after Valsalva, suggesting that a right-to-left shunting occurred and that PFO may contribute to hypoxaemia . The methodology of this study was criticized, regarding patient selection and exclusion criteria .
In the second study, Beelke et al. studied 78 patients with OSA and 89 controls using transcranial Doppler . All subjects with OSA had AHI > 10 during polysomnography, and the mean ± SD AHI was 52 ± 25 . PFO was found in 21/78 patients with OSA (26.9%) versus 13/89 controls (14.6%; P < 0.05) . Among patients with OSA, 85% only presented PFO during Valsalva.
Considering these two studies, it is possible to believe that the prevalence of PFO might be higher in OSA patients. However, in the second study , the percentage of patients with OSA and PFO (26.9%) is close to the percentage found in the general population in other epidemiological studies. The small size of these studies, the difference of patient selection, and the technique of detection is clearly insufficient to demonstrate a direct correlation between these two pathologies, and larger epidemiological studies are now required.
Moreover, these studies only show an association rather than a causal relation, and do not answer the questions of whether OSA increases the risk of PFO or whether PFO is a risk factor itself for OSA. Furthermore, no precise data on stroke before or after the diagnosis of OSA and PFO were presented in these studies.
Pathophysiological interactions between obstructive sleep apnoea and patent foramen ovale
Several pathophysiological interactions exist between OSA and PFO ( Fig. 1 ). OSA is the consequence of repetitive pharyngeal collapse while diaphragmatic efforts are maintained. At first, an obstructive apnoea may be assimilated to a Muller manoeuvre (inspiration effort against a closed upper airway), which can provoke severe negative intrathoracic pressure as low as −80 cm H 2 0 . If the obstruction persists, alternating sequences of Muller and Valsalva manoeuvres may be undertaken, leading to important intrathoracic pressure variations. Enhanced venous return is observed when intrathoracic pressure decreases, leading to repetitive transient pressure and volume elevation in the right heart, predisposing to right-to-left shunting . Shiomi et al. undertook TTEs during the sleep of apnoeic patients, and observed a leftward shift of the interventricular septum during the apnoea in half of their patients, all presenting the lowest intrathoracic pressure during the apnoea periods . This shift was correlated with a decrease in arterial blood pressure during inspiration, called a pulsus paradoxus (as described in cardiac tamponade), explained through an interdependence mechanism limiting left ventricular filling . Pulsus paradoxus was more frequent in younger people who could develop more negative inspiratory oesophageal pressure . Interestingly, CPAP limits the variations of intrathoracic pressure and abolishes these interdependence mechanisms and the pulsus paradoxus .
