Author + information
- Received April 4, 2016
- Revision received June 6, 2016
- Accepted June 30, 2016
- Published online September 26, 2016.
- Ashish H. Shah, MD, MD-Researcha,
- Mark Osten, MDa,
- Andrew Leventhal, MD, PhDa,
- Yvonne Bach, BScHa,
- Daniel Yoo, BScHa,
- Danny Mansour, BScHa,
- Lee Benson, MDb,
- William M. Wilson, MDa and
- Eric Horlick, MDCMa,∗ ()
- aPeter Munk Cardiac Centre and Toronto Congenital Cardiac Centre for Adults, Toronto General Hospital, University Health Network
- bThe Labatt family Heart Centre, The Hospital for Sick Children, Division of Cardiology, The University of Toronto School of Medicine, Toronto, Ontario, Canada
- ↵∗Reprint Requests and Correspondence:
Dr. Eric Horlick, Toronto General Hospital, Room 6E-249, Toronto, Ontario M5G 2C4, Canada.
Objectives This study reviewed a series of patients treated with transcatheter closure of septal defect to treat platypnea-orthodeoxia syndrome, with specific attention to septal characteristics and device choice.
Background Platypnea–orthodeoxia syndrome is an uncommon condition characterized by positional dyspnea and hypoxemia due to intracardiac right-to-left shunting through a patent foramen ovale (PFO), an atrial septal defect, or pulmonary arteriovenous malformations. Percutaneous closure of such defects is the treatment of choice.
Methods In this single-center series, 52 patients were treated with percutaneous closure of an interatrial communication after presentation between January 1997 and July 2015. Septal morphology, clinical, procedural, and outcomes data were analyzed.
Results All patients had a PFO; however, nearly one-quarter required a non-PFO device (11 Amplatzer Septal Occluder and 1 post-infarct muscular VSD), as opposed to a dedicated PFO device to achieve shunt occlusion. These patients were characterized by an aneurysmal septum, shorter primum septum overlap with the secundum septum, and greater septal angulation from the midline. After closure, all demonstrated acute improvements in oxygen saturation (pre-procedure: 81 ± 8%; post-procedure: 95.1 ± 0.5% on room air). Each patient was treated with a single device and no one required re-intervention.
Conclusions Patients presenting with platypnea–orthodeoxia syndrome can be treated successfully with a percutaneous intervention often requiring a variety of devices. Those requiring a non-PFO–type device had a greater prevalence of an aneurysmal septum, shorter primum septal overlap with the secundum septum, and greater septal angulation with the midline.
Platypnea–orthodeoxia syndrome (POS) is an uncommon condition characterized by dyspnea and objective evidence of hypoxemia in the upright that resolves in the supine position (1). Shunting occurs through a patent foramen ovale (PFO) in the majority of patients, an atrial septal defect (ASD) in a minority (2), and rarely in patients with pulmonary arteriovenous malformations (3). A number of conditions and surgical procedures that distort the usual cardiac position and orientation have been associated with the syndrome including cardiothoracic or abdominal surgeries and aortic aneurysms (4–17). Although PFOs may be found in ≤25% of the general population, the occurrence of POS is distinctly uncommon (Table 1) (4–17).
One proposed mechanism for POS is interatrial septal (IAS) distortion and a change in the septal relationship to the inferior vena cava (IVC), resulting in streaming of the IVC blood through the septal defect (16). As such, the presentation can be unrelated directly to cavity pressures, and can have an acute onset, resolving quickly with percutaneous closure (4). Generally, percutaneous intervention is technically uncomplicated, effective, and free of significant complications (18,19). However, there is limited information available with regard to the septal characteristics and anatomy, as well as guidance as to the types of devices that are optimal for these anatomies.
Materials and Methods
Between January 1997 and July 2015, 938 patients underwent percutaneous closure of a PFO at the Peter Munk Cardiac Centre at Toronto General Hospital, Canada. The indication for closure was POS in 52 patients. Excluded in this review were patients with an Ebstein anomaly. Demographics and procedural data were collected from the prospective data registry maintained at the institution. The local research and ethics board approved the study.
Available transesophageal echocardiograms (TEE) were reviewed (n = 44). An atrial septal aneurysm was defined if septal excursion was >15 mm, whereas 10 to 15 mm of excursion was described as mobile (20). The length of membranous primum septal overlap with the superior limbic body of the secundum septum was measured in the aortic short axis view on TEE (Figure 1A), as was the presence or absence of a Eustachian valve, right atrial membrane, or Chiari network. Post-procedural transthoracic echocardiograms with and without bubble studies were reviewed to determine procedural success and presence of a residual shunt (n = 34). The studies were performed both at rest and after a Valsalva maneuver. A residual shunt was defined if >3 bubbles were visualized in the left atrium within 3 cardiac cycles after their appearance in the right atrium (RA). Passage of 5 to 10, 10 to 20, and >20 bubbles defined small, medium, and large residual shunts, respectively (21).
Procedural success was defined as complete or near-complete occlusion of the defect and resultant improvement in oxygen saturation (>93% in room air or with O2 in the presence of underlying lung pathology) without postural change.
Percutaneous interatrial communication closure
Informed consent was obtained and the procedures performed with local anesthesia and appropriate use of intravenous sedation. Intravenous unfractionated heparin (100 U/kg) was administered adjusted to an activated clotting time of >275 s. Right heart catheterization and hemodynamic measurements were performed. We also measured oxygen saturation in individual pulmonary veins to distinguish between global and regional lung dysfunction versus A right-to-left shunt (RLS). An Amplatzer PFO (St. Jude Medical, St. Paul, Minnesota) device was initially chosen in all patients; a 35-mm device used in those with an aneurysmal septum, and a 25-mm device in the remainder. After device deployment a residual leak was assessed by: 1) a contrast injection through the delivery sheath in the RA; 2) intracardiac echocardiography color Doppler (n = 35, 63.5%) (there was a trend to use more intracardiac echocardiography as the series progressed); and 3) an increase in arterial oxygen saturation. Subjectively, if there was a moderate to severe residual shunt after device deployment, the PFO device was removed, balloon sizing of the defect was performed, and an appropriately sized non-PFO device was chosen. Patient charts were reviewed and complications assessed, specifically, vascular, bleeding, need for urgent surgery, stroke or transient ischemic events, arrhythmias, myocardial infarction, and device embolization or erosion.
Assessment of septal distortion
Given the 3-dimentional structure of the IAS, it is not possible to assess fully the relationship of the defect to the IVC. As a proxy, we measured the angulation between the device (discs) and the midline from the coronal view of post-device deployment computed tomography (CT) scans from 14 patients (Figure 2). From 886 patients who underwent PFO closure for non-POS indications, the clinical charts from the latest consecutive 290 patients were reviewed and 17 were identified who had post-device cardiac CT scans, used as a comparison group. As a control group, we also reviewed the clinical charts of 250 consecutive patients who underwent percutaneous closure of a secundum ASD with an Amplatzer septal occluder (ASO) device, and identified 14 patients who also had cardiac CT scans after ASO deployment. Similar measurements on the disc angulation with the midline were performed.
Follow-up was performed between 2 and 3 months after the procedure and consisted of a clinical history and physical examination with pulse oximetry in the supine and standing positions, electrocardiography, and transthoracic echocardiography with a bubble study.
Values are presented descriptively as numbers, percentages, frequency (%) or mean ± SD. Due to small number of cases, comparison between continuous variables was not performed.
Patient demographics, symptoms, oxygen saturations, cardiovascular risk factors, and routine blood test results are reported in Table 2. Twenty-four patients were male; the mean age was 65.6 ± 15.1 years of age, and body mass index 29.1 ± 8.0 kg/m2. All patients were dyspneic with a mean New York Heart Association functional class score of 2.6 ± 0.8; more than one-half were in New York Heart Association functional class III/IV. Five patients (9.6%) had cyanosis at rest in the supine position. The average saturation in the supine position was 93 ± 3% versus 81 ± 8% in the upright position; 19 patients (36.5%) were on oxygen when these recordings were made. One patient had a history of a transient ischemic attack and another had a stroke several years before this intervention. The associated medical conditions are listed in Table 3. Pneumonectomy was the most common surgical procedure (n = 10, 19.2%), involving the right lung in 9 patients. Other pulmonary pathologies (chronic obstructive pulmonary disease, obstructive sleep apnea, pulmonary metastasis, and cavitating lung lesions) were present in 7, and ascending aortic dilation or preceding aortic root surgery was noted in 8 patients. Patients were referred from 4 days to 5 years after surgery. Three patients had surgically palliated congenital heart disease. Four patients presented after a right ventricular infarction, and 1 had symptoms after a percutaneous coronary intervention to the right coronary artery. Interestingly, 2 patients demonstrated postural hypoxemia after hip and knee replacement surgeries during their in-hospital stays. In 20 patients (38.5%), no associated condition was identified. Those patients without a clear pre-disposing event were relatively older (69.7 ± 15.8 vs. 63.1 ± 14.4 years of age), more likely to be a female (75% vs. 40.6%), and were symptomatic for longer durations (150 ± 156 weeks vs. 61.9 ± 109 weeks) in comparison with those with an identifiable associated condition.
The average duration of symptoms before percutaneous intervention was 101 ± 133 weeks, ranging from 3 days to 10 years. Two of the very late presenting patients were undergoing phlebotomies to treat erythrocytosis-related hyperviscosity symptoms. The average hemoglobin for all patients was 140.8 ± 4.0 g/l (range 83 to 193 g/l), of which 18 patients had a hemoglobin of >150 g/l (average 168 ± 3). These patients with a hemoglobin of >150 g/l had a significant delay from symptom onset to diagnosis of POS (188 ± 155 weeks vs. 49 ± 115 weeks) (Table 2).
Pre-procedural TEE images were available for review in 44 patients, with a PFO the culprit in all. Review of the clinical and procedural notes from the other 8 patients also described a PFO seen on TEE or intracardiac echocardiography. Color-flow Doppler demonstrated bidirectional flow in 38 patients (73.1%), and thus bubble studies were not performed; whereas in the remainder a large RLS was confirmed by bubble study. An aneurysmal septum was observed in 65.9%, and in one-quarter of the patients there was a prominent Eustachian valve or Chiari network. An aneurysmal septum was more likely to be present in patients who required a non-PFO device (11 of 12) to achieve adequate closure compared to those in whom closure was achieved with a PFO device (18 of 32). The membranous primum septum overlapped less with the superior limbic body of the secundum septum in patients who required a non-PFO device to close the defect, 0.6 ± 0.1 cm (n = 6) versus 1.3 ± 0.1 cm (n = 24). Differences in septal characteristics and angulation between patients, who were treated with either a PFO or Amplatzer non-PFO device, are described in Table 4.
The POS constituted 5.5% of the total cohort of patients undergoing PFO closure during the study period. Pre-intervention right heart hemodynamics revealed normal pulmonary artery pressures in all patients (Table 5). Measurement of pulmonary venous oxygen saturation helped to differentiate patients with atrial level RLS versus those with a mixed picture of RLS with intrinsic lung disease as the cause of hypoxia. Transcatheter device closure was successfully performed in all patients. Five different devices were used (Table 5). In 2 patients, an Amplatzer PFO device was used to close the defect after initial failure of a CardioSEAL device. In those with a moderate to severe residual shunt after initial Amplatzer PFO device deployment, the device was removed, balloon sizing of the defect was performed, and a self-centering Amplatzer non-PFO device was used (n = 12; 11 ASO and 1 post-infarct muscular VSD occluder), the size ranging from 12 to 30 mm. A total of 66 devices were deployed to close the 52 defects with a single device deployed in each defect. All patients demonstrated marked improvement in their oxygen saturations immediately after device deployment (pre-procedure: 81 ± 8%; postprocedure: 95.1 ± 0.5% on room air). Of the 886 patients who underwent PFO closure for non-POS indications, none required a non-PFO device implantation for complete occlusion.
Procedural complications occurred in 7.7% (n = 4). One patient had air embolization to the right coronary artery and transient ST-segment elevation followed by ventricular fibrillation treated with coronary catheter–guided aspiration of air and DC cardioversion. The patient made an uneventful recovery. This was thought to be due to paradoxical embolization of air through an intravenous line without a bubble trap. Three patients had atrial fibrillation during the procedure; 1 spontaneously reverted to sinus rhythm, and the other 2 required electrical cardioversion in the catheter laboratory. After intervention, 2 patients died during their concurrent hospital admission due to non–procedure-related medical conditions. One patient died of metastatic small cell lung cancer, and another patient of sepsis after a pneumonectomy to treat metastatic colon cancer.
Angulation of the right atrial disc in patients with POS requiring an ASO device was 69 ± 3°; compared to 43 ± 7° in defects closed with a PFO device. The PFO device angulation in POS was similar to the 17 patients who underwent PFO closure for a non-POS indication (46 ± 3°), or the patients who underwent a secundum ASD closure using an Amplatzer ASO device (47 ± 3°). As such, the ASO devices were more angulated in patients with POS in comparison with those with secundum ASD, very likely representing the septal distortion rather than the device itself. Detailed description of septal characteristic and device angulation is described in Table 4.
Late follow-up was available in 35 patients (67.3%) at a mean 26.4 ± 15 months. Four patients had mild exertional hypoxemia with saturations ranging from 88 to 92% in room air that was thought to be due to underlying pulmonary disease, and no residual shunt was found. Overall, a small residual shunt was identified by bubble study in 19.6% (n = 11). In the non-PFO device group, bubble contrast studies demonstrated a small residual shunt in 2 patients, 1 at rest the other after a Valsalva maneuver only. In the PFO device group, 9 patients had a positive bubble study demonstrating a small residual shunt. All the patients were asymptomatic; their O2 saturation was normal without a postural change, and no repeat interventions were required. During late follow-up, 2 additional patients developed atrial fibrillation 12 and 15 months after the intervention.
POS is an uncommon condition. All of our patients had a PFO as the underlying interatrial communication, which constituted 5.5% of all patients undergoing PFO closure at our institution during the study period. The most commonly associated conditions included pneumonectomy (often the right lung) (Figure 3), ascending aortic dilation, and arch surgery (Figures 4 and 5), as reported previously (4,7,16). Interestingly, 20 patients in our series (38.5%) did not have an identifiable cause. A prolonged delay before the formal diagnosis was not uncommon (16). The diagnosis was made earlier in those who experienced symptoms immediately after surgery, usually during their hospital stay. Patients presenting late after symptom onset, demonstrated appropriate erythrocytosis with a resultant increase in hematocrit and a physiological response to chronic hypoxia (22).
In the present series, an aneurysmal septum was a common finding in those with POS. Septal aneurysms have previously been reported to be an independent risk factor associated with recurrent stroke and paradoxical embolism through a PFO, suggestive of its affiliation with exaggerated RLS (23). TEE also demonstrated the presence of bidirectional flow in more than two-thirds of the patients. This is an important finding; although these patients demonstrate hypoxemia mainly in the upright position, they also have persistent atrial level RLS in the supine position, a finding confirmed by angiography in the IVC and improvement in oxygen saturation by balloon/device occlusion of the defect, when in the supine position.
It is imperative to measure oxygen saturation in individual pulmonary veins to distinguish between lung or lobar functional abnormalities versus atrial level RLS, especially in patients who have concomitant lung/pleural disease or have lung collapse/atelectasis post-operatively. Similarly, it is important to exclude pulmonary embolism in these patients, before proceeding with transcatheter interventions. The use of pulmonary venous saturations offers a good estimation of regional lung function.
Two theories have been proposed as etiologies for RLS in POS. The first is hemodynamic based on the pressure gradient between the atria, as we noted in one of our patients after an RV infarction. This is, however, a relatively rare phenomenon (24,25) and not present in the majority of patients with POS subjected to direct hemodynamic measurement. The second proposed mechanism is based on altered flow dynamics and streaming of blood from the right to left atrium. To this end, 3-dimensional, velocity-encoded, phase contrast 4-dimensional magnetic resonance imaging of the right and the left atrium will help us to understand blood flow vortex alteration in patients presenting with POS (26).
Because a PFO is present at birth, we and others hypothesize that the late presentation of POS can be explained by age-related or post-operative anatomic distortion of the heart resulting in an altered relationship between the interatrial septum, the IVC, and possibly the superior vena cava, affecting the flow pattern into the RA. Normally, the blood entering the RA from the cava stream, with flow from the superior cava deflected anteriorly toward the tricuspid valve (crista terminalis aligning the blood) whereas IVC flow is directed posterosuperiorly toward the upper part of the IAS (Eustachian valve aligning flow), before turning downward toward the tricuspid valve (27). A combination of these flow patterns creates a single uniform and well-organized clockwise vortex in the RA, when viewed from the right side of the patient (28). The fluid motion in the vortex exerts dynamic pressure in addition to the hydrostatic pressure, lowest in the central core region, which increases as the distance from the axis of the vortex increases. Precipitating conditions in these patients can result in horizontal distortion of the heart, the RA, and the interatrial septum. These cumulative changes influence vortex formation, as noted in patients with an ASD (29) or repaired tetralogy of Fallot (28). Such a change in vortex formation with the higher dynamic pressure in the outer layer can result in a differential pressure gradient across the interatrial septum that can result in RLS. The invasive hemodynamic pressure measurements performed in the cardiac catheter laboratory are taken from the center of the cavity and as such cannot localize precisely the pressure gradients associated with the intracardiac vortices. Although the mean left atrial pressure is higher than in the RA, simultaneous transcatheter pressure measurements have demonstrated intermittent changes in pressure gradient (higher RA pressure than LA), especially with inspiration (30) or during early systole (16). As such, one can observe preferential streaming of blood from the right to left atrium during contrast injection in the IVC, as reported by others (16).
One of the important observations in our study was the need of a non-PFO device to achieve complete closure of a PFO in 12 patients (23.1%). We had hypothesized that, given that almost all those presenting for PFO closure could be treated with an Amplatzer PFO occluder, this would be a sensible first choice of device to implant. Predictors of a non-PFO device use include shorter membranous primum septal overlap with the superior limbic body, aneurysmal interatrial septum, and higher septal angulation to the midline. These findings may describe a subgroup of patients in whom the interatrial septum was stretched necessitating the use of a device with a larger waist (e.g., ASO) to occlude the defect. Godart et al. (16) also made similar observations of horizontal displacement of the interatrial septum that can expose the stretched defect in line with the IVC flow. Interestingly, one of their patients also required an ASO device to achieve complete closure of the PFO. In contrast with multiple case series that reported a need for repeat intervention due to residual shunt or even persistence of POS, each of our patients were treated with a single device to achieve successful closure of the defect (4,9,31). Although 20% of our patients had trace/mild residual shunt, their oxygen saturation was in the normal range and none required repeat intervention.
Although 19.6% of patients demonstrated a small residual shunt after the intervention, all patients had improvements in their oxygen saturations. Four patients reported persistent exertional dyspnea, without evidence of residual RLS. The symptoms were thought to be due to an underlying pulmonary pathology.
This was a nonrandomized, retrospective observational study with its inherent limitations. Many of the patients were investigated by TEE at their referring hospital, and some of the acquired images were of suboptimal quality, limiting our ability to define the IAS anatomy. After the intervention, a small number of patients were investigated by CT scan, and therefore the septal angulation measurement was limited to these patients only. Altered vortex formation is a hypothetical, albeit possible, mechanism resulting in RLS. However, it is not possible to prove or disprove its existence with imaging modalities available as part of this retrospective study. Late follow-up was limited to two-thirds the cohort and hindered by remote location of many patients.
The POS is an uncommon condition that is likely to be underdiagnosed and perhaps the diagnosis is considered late from symptom onset. Its presentation in older patients underscores the need for a high index of suspicion. Altered anatomy-mediated flow dynamics is very likely to be the underlying pathophysiology. Patients with angulated, aneurysmal septa with shorter septal overlap are likely to require a non-PFO device to achieve optimal defect closure. Percutaneous closure of the PFO with a variety of devices is technically feasible, with a low complication rate, and significant clinical impact.
WHAT IS KNOWN? RLS of deoxygenated blood, mainly through either a PFO or an ASD results in POS. Percutaneous closure is the treatment of choice. However, a proportion of patients continue to have either residual shunt or require repeat intervention.
WHAT IS NEW? For the first time, we describe the septal characteristics of patients presenting with POS, and their correlation with device selection to offer complete occlusion of the PFO. All of our patients were treated with single device, and none required repeat intervention. A septum that is aneurysmal, angulated, and has a shorter overlap with the septum secundum requires a non-PFO device to achieve optimal defect closure. Ours is the second largest series that adds valuable experience in the world literature.
WHAT IS NEXT? Assessment and comparison of blood flow vortex alteration by 4-dimensional MRI of the right and left atrium in patients with and without POS presentation.
Dr. Horlick is supported by the Peter Munk Chair in Structural Heart Disease Intervention.
Dr. Horlick is a proctor and consultant for St. Jude Medical. The structural heart disease program at the Toronto General Hospital receives support for its educational and research missions from St Jude Medical. All other authors have reported that they no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- atrial septal defect
- Amplatzer septal occluder
- computed tomography
- interatrial septum/septal
- inferior vena cava
- patent foramen ovale
- platypnea–orthodeoxia syndrome
- right atrium
- right-to-left shunt
- transesophageal echocardiogram
- Received April 4, 2016.
- Revision received June 6, 2016.
- Accepted June 30, 2016.
- American College of Cardiology Foundation
- Seward J.B.,
- Hayes D.L.,
- Smith H.C.,
- et al.
- Kumar N.,
- Kraemer R.R.,
- Murthy R.K.,
- Hartig J.R.
- Takaya Y.,
- Akagi T.,
- Kijima Y.,
- et al.
- Godart F.,
- Rey C.,
- Prat A.,
- et al.
- Mojadidi M.K.,
- Gevorgyan R.,
- Noureddin N.,
- Tobis J.M.
- Homma S.,
- Di Tullio M.R.,
- Sacco R.L.,
- Sciacca R.R.,
- Smith C.,
- Mohr J.P.
- Pearson A.C.,
- Nagelhout D.,
- Castello R.,
- Gomez C.R.,
- Labovitz A.J.
- Stewart M.J.
- Cabanes L.,
- Mas J.L.,
- Cohen A.,
- et al.
- Burke M.A.,
- Mikati I.A.
- Essop M.R.,
- Essop A.R.,
- Bedhesi S.,
- Sareli P.E.
- Markl M.,
- Carr M.,
- Ng J.,
- et al.
- Francois C.J.,
- Srinivasan S.,
- Schiebler M.L.,
- et al.
- Shiraishi Y.,
- Hakuno D.,
- Isoda K.,
- Miyazaki K.,
- Adachi T.
- Knapper J.T.,
- Schultz J.,
- Das G.,
- Sperling L.S.