Author + information
- Received January 20, 1999
- Revision received June 14, 1999
- Accepted August 5, 1999
- Published online November 15, 1999.
- Harald P Kühl, MDa,* (, )
- Rainer Hoffmann, MDa,
- Marc W Merx, MDa,
- Andreas Franke, MDa,
- Christof Klötzsch, MD∗,
- Wolfgang Lepper, MDa,
- Thorsten Reineke, BSc†,
- Johannes Noth, MD∗ and
- Peter Hanrath, MD, FACCa
- ↵*Reprint requests and correspondence: Dr. Harald P. Kühl, Medizinische Klinik I, Universitätsklinikum der RWTH, Pauwelstrasse 30, 52057 Aachen, Germany
We sought to evaluate whether transthoracic contrast echocardiography using second harmonic imaging (SHI) is a diagnostic alternative to transesophageal contrast echocardiography (TEE) for the detection of atrial right to left shunt.
Paradoxic embolism is considered to be the major cause of cerebral ischemic events in young patients. Contrast echocardiography using TEE has proven to be superior to transthoracic echocardiography (TTE) for the detection of atrial shunting; SHI is a new imaging modality that enhances the visualization of echocardiographic contrast agents.
We evaluated 111 patients with an ischemic cerebral embolic event for the presence of atrial right to left shunt using an intravenous (IV) contrast agent in combination with three different echocardiographic imaging modalities: 1) TTE using fundamental imaging (FI); 2) TTE using SHI; and 3) TEE. The severity of atrial shunting and the duration of contrast visibility within the left heart chambers were evaluated for each imaging modality. Image quality was assessed separately for each modality by semiquantitative scoring (0 = poor to 3 = excellent). Presence of atrial right to left shunt was defined as detection of contrast bubbles in the left atrium within the first three cardiac cycles after contrast appearance in the right atrium either spontaneously or after the Valsalva maneuver.
A total of 57 patients showed evidence of atrial right to left shunt with either imaging modality. Fifty-one studies were positive with TEE, 52 studies were positive with SHI, and 32 were positive with FI (p < 0.001 for FI vs. SHI and TEE). The severity of contrast passage was significantly larger using SHI (61.6 ± 80.2 bubbles) compared to FI (53.7 ± 69.6 bubbles; p < 0.005 vs. SHI) but was not different compared to TEE (43.9 ± 54.3 bubbles; p = NS vs. SHI). The duration of contrast visibility was significantly longer for SHI (17.4 ± 12.4 s) compared to FI (13.1 ± 9.7 s; p < 0.001) and TEE (11.9 ± 9.6 s; p < 0.02). Mean image quality improved significantly from FI (1.5 ± 0.8) to SHI (2.0 ± 0.8; p < 0.001 vs. FI) and TEE (2.5 ± 0.7; p < 0.001 vs. SHI).
In combination with IV contrast injections, TEE and SHI have a comparable yield for the detection of atrial right to left shunt. Both modalities may miss patients with atrial shunting. In young patients with an unexplained cerebrovascular event and no clinical evidence of cardiac disease, a positive SHI study may obviate the need to perform a TEE study to search for cardiac sources of emboli.
Patent foramen ovale (PFO) with atrial right to left shunt is the most frequent cardiac finding in patients <55 years of age with an otherwise unexplained ischemic cerebral insult (1,2). A causative relationship among venous thrombosis, PFO, and ischemic neurologic events has been suggested in several studies (3–5). Thus, an imaging modality that reliably detects an atrial right to left shunt is required. An evolution of methods has occurred for direct and indirect detection of PFO from transthoracic echocardiography (TTE) using contrast agents (6)to transesophageal echocardiography (TEE) using color Doppler (7)as well as contrast agents (8)and transcranial Doppler ultrasound in combination with contrast agents (9,10). Contrast echocardiography (CE) has evolved as the diagnostic method of choice for the assessment of PFO owing to the direct visualization of an atrial right to left shunt (11). Transesophageal CE is superior to TTE for the detection of PFO because of the improved image quality allowing superior visualization of the atria, atrial septum and the site of contrast passage (12,13). However, false negative results have been reported using this imaging modality in comparison to transcranial Doppler ultrasound (14,15). This finding has been attributed to the limited ability of patients to perform the Valsalva maneuver after sedation and esophageal intubation.
Second harmonic imaging (SHI) is a new imaging modality that is based on the principle of receiving double the emitted ultrasound frequency. It has shown to improve the visualization of left heart echo contrast agents and to improve transthoracic two-dimensional (2D) image quality (16,17). This study was designed to evaluate TTE using SHI as a diagnostic alternative to TEE for the detection of atrial right to left shunt in patients with cerebral embolic events.
Baseline characteristics of the study population are given in Table 1. To evaluate the echocardiographic findings dependent on patient age, patients were divided into two age groups, those <55 years (group I) and those ≥55 years of age (group II). The study population consisted of 111 consecutive patients (mean age 55.9 ± 15.2 years, range 20 to 86 years) referred from the Department of Neurology to our echocardiography laboratory for the evaluation of cardiac sources of embolism. Only patients with obviously ischemic neurological deficits were included into the study. A transient ischemic attack was assumed if a patient presented with symptoms compatible with a cerebral ischemia and that resolved completely within 24 h. Patients with a history of migraine and headaches after the development of transient neurological deficits were excluded. Patients with significant carotid artery disease and abnormal transcranial Doppler findings or patients who showed apparently lacunar syndromes (e.g., pure sensory stroke or pure motor stoke) or revealed a new lacunar infarction on computed tomography (CT) or magnetic resonance imaging (MRI) scanning were not referred for echocardiography. Patient chart review was used for the evaluation of the clinical history and assessment of atrial fibrillation by electrocardiogram (ECG) and Holter monitoring. All patients gave informed consent.
All studies were performed either with a Sonos 5500 (HP, Andover, Massachusetts) or a Sequoia 256 system (Acuson, Mountain View, California). Both machines use broad-band transducers capable of SHI (Hewlett-Packard S4 with 1.8/3.6 MHz or 2.1/4.2 MHz and Acuson Sequoia 3 VS with 1.75/3.5 MHz).
The patients underwent a complete TEE study including color Doppler using a multiplane probe (Omni II, HP, Andover, Massachusetts). The TEE study was performed with the patient positioned in the left lateral decubitus position after topical anesthesia of the pharynx with lidocaine. All patients were mildly sedated with intravenous (IV) administration of 2 to 3 mg of midazolam. For the purpose of this study, contrast injections were performed in the transesophageal “four-chamber view” in the 0° image plane of the transducer. Care was taken to visualize optimally the left atrium, the left ventricle and the interatrial septum. Additional contrast injections were performed in 40° to 60° image planes and/or 110° to 130° image planes as needed to demonstrate clearly the site of contrast passage through the interatrial septum.
The TTE studies were performed following the TEE study at least 3 min after contrast bubbles had disappeared from right heart chambers. There was no change in the patient position between the TEE and TTE studies. For the transthoracic CE study an apical four-chamber view with optimal delineation of both atria, the interatrial septum and both ventricles was selected. Transthoracic echocardiographic images were acquired in the fundamental imaging mode using the highest possible transducer frequency that still allowed clear delineation of the cardiac morphology. Gain settings were adjusted individually for each patient to optimally visualize myocardial and valvular structures and the interatrial septum. After contrast bubbles had completely cleared from the right heart chambers the transducer was switched into the harmonic mode, holding the probe in the same position as in the fundamental imaging mode. In this position the receiver gain was again adjusted individually in each patient to optimally delineate myocardial and valvular structures and the interatrial septum, and the mechanical index was set to maximum. Thereafter, contrast injections were repeated. All echocardiographic studies were recorded on S-VHS videotape for off-line analysis.
The CE was performed using a polygelatin solution (Gelifundol, Biotest Pharma, Dreieich, Germany). The contrast medium was prepared by shaking the vial and repeatedly drawing up and reinjecting 10 ml of the solution into the vial with a 10-ml syringe until the clear solution became cloudy and foamy. Because of their large size the generated air-filled microbubbles are not capable of passing the pulmonary capillary bed. After extruding all microscopic air, 10 ml of the contrast solution was rapidly injected into a right antecubital vein through an 18-gauge venous canula. At least two contrast injections were performed in each study (i.e., the transesophageal study and the transthoracic study using the fundamental as well as second harmonic mode). Thus, in each patient at least six injections of the contrast solution were performed. Immediately after the contrast bubbles had reached the right atrium, each patient was asked to perform the Valsalva maneuver. Echocardiographic images were recorded until the contrast bubbles had completely disappeared from the left heart chambers.
All studies were analyzed off-line from the videotapes. To account for possible bias the different TTE studies from each patient using either the fundamental or second harmonic mode were copied in random order from the original videotapes to a single S-VHS videotape. Each study was assigned a number that was used for patient and modality identification after data analysis. Therefore, the observers were completely blinded to the imaging mode of the video sequences and to the patient data by obscuring the display of the technical instrument settings and patient data. The videotapes were analyzed by two experienced observers (HPK, MWM). In case of uncertainty or divergent results, the studies were reviewed by a third observer (RH) and consensus interpretation was achieved. All TEE and TTE studies were analyzed for the presence of atrial shunting. An atrial right to left shunt was diagnosed when ≥5 bubbles appeared in the left atrium within the first three cardiac cycles after contrast appearance in the right atrium either spontaneously or after the Valsalva maneuver. Bubbles appearing in the left atrium after this period were classified as pulmonary shunt. Quantitative analysis of the video data consisted of the following:
1. Severity of contrast passage.The stop-frame image showing the most intensive contrast filling of the left heart chambers was selected for analysis. In this image frame the contrast bubbles were counted directly in the left atrium and left ventricle from the video screen up to a bubble number of 150.
2. Duration of contrast passage.The duration of contrast visualization was defined as the time period that elapsed between the first and the last video frame showing ≥5 contrast bubbles within the left atrium and left ventricle.
3. Image quality.For all three echocardiographic imaging modalities 2D image quality was graded semiquantitatively. Considering the visibility of endocardial borders, valvular structures and the interatrial septum, four different levels were defined thus:
0 = poor image quality:no clear visualization of endocardial borders, valvular structures and the interatrial septum;
1 = moderate image quality:failure to detect endocardial borders in >2 segments according to the segmentation of the American Society of Echocardiography (ASE) (18), hazy delineation of valve structures or no clear delineation of the atrial septum due to clutter;
2 = good image quality:failure to detect endocardial borders ≤2 segments, adequate visualization of valve structures and atrial septum with minimal clutter;
3 = excellent image quality:complete visualization of endocardial borders, valves and the interatrial septum without clutter.
Numerical results are given as mean ± standard deviation. The McNemar test was applied to compare dichotomous variables. For the comparison of percentages and proportions the Fisher exact test was used. The Wilcoxon paired test was used to compare differences in mean image quality, severity and duration of contrast opacification of the left heart chambers between the different modalities. All tests were two-tailed and a p value of <0.05 was considered significant.
Echocardiographic findings by TEE
The pathomorphological results detected by TEE for the entire study population and divided into age groups are given in Table 2. Aortic plaques were significantly more often detected in older patients (p < 0.005). Although the prevalence of spontaneous echocardiographic, left atrial appendage thrombus, vegetations, depressed left ventricular (LV) function and atrial fibrillation was not different between the two age groups because of the low patient numbers, the combined prevalence of these findings was significantly different between patients <55 years versus those ≥55 years of age (0 vs. 13; p = 0.002).
Detection of atrial right to left shunt by contrast echocardiography
Of the 111 patients, 57 (51.3%) showed evidence of atrial right to left shunt with either imaging modality. Using venous contrast injections an atrial right to left shunt could be demonstrated with fundamental imaging (FI) in 32 patients (28.8%), with SHI in 52 patients (46.8%), and with TEE in 51 patients (45.9%; Fig. 1). No significant difference existed between TEE and SHI for the detection of atrial right to left shunt. In contrast, results were significantly different between TEE and FI (p < 0.001) and between SHI and FI (p < 0.001). All patients with a positive FI study also showed evidence of atrial shunting using SHI (100%). Of the 51 patients with an atrial right to left shunt detected by TEE, results were also positive in 31 patients using FI (60.8%) and in 46 patients using SHI (90.2%; p < 0.01 for FI vs. SHI; Fig. 2). A negativeTEE study was found in 60 patients. However, in these patients FI was positive in one patient (1.7%) and SHI was positive in six patients (10.0%; p = NS for SHI vs. FI; Fig. 2). Thus, a right to left shunt by CE could be demonstrated in six patients only by TTE using SHI (including one patient who also showed a positive result with FI) and in five patients only by TEE.
The echocardiographic findings in patients who demonstrated an atrial right to left shunt only with SHI included an isolated PFO in five patients and an atrial septal aneurysm as proposed by Hanley (19)with PFO in one patient. The five patients who were only positive with TEE showed an isolated PFO in four patients and an atrial septal aneurysm with PFO in one patient. Figure 3shows an example of a patient with an atrial right to left shunt detected only by SHI.
Severity and duration of atrial right to left shunt
The severity of contrast passage was assessed as significantly larger by SHI (61.6 ± 80.2 bubbles) as compared to FI (53.7 ± 69.6 bubbles; p < 0.005) and was not different between TEE (43.9 ± 54.3 bubbles) and either SHI or FI (Fig. 4). The duration of contrast visualization of left heart chambers was assessed as significantly longer by SHI (17.4 ± 12.4 s) as compared to TEE (11.9 ± 9.6 s; p < 0.02) and FI (13.1 ± 9.7 s; p < 0.001). The difference was not statistically significant for the comparison between TEE and FI (Fig. 4). In patients with an atrial right to left shunt detected exclusively either by SHI (n = 6) or by TEE (n = 5) shunt severity was only low. The absolute number of bubbles was not significantly different between the studies in which the atrial shunt could only be detected by SHI or only by TEE (21.8 ± 10.0 vs. 29.8 ± 29.0 bubbles, p = NS) and no significant difference in the duration of atrial right to left shunt was noticed (19.2 ± 9.5 vs. 9.2 ± 7.5 s, p = NS). No patient with severe shunting (>80 bubbles) as assessed by either TEE or SHI was missed by the other imaging modality. Considering only patients with moderate contrast passage (30 to 80 bubbles) two patients with a positive TEE finding were not identified by SHI, and one patient with a positive SHI study was not detected by TEE. Four patients with moderate or severe shunting by TEE or SHI were missed by FI.
Impact of the imaging modality on image quality
Mean image quality was 1.5 ± 0.8 for FI, 2.0 ± 0.8 for SHI and 2.5 ± 0.7 for TEE and was significantly different among the three imaging modalities (p < 0.001 for FI vs. SHI, SHI vs. TEE, and FI vs. TEE). The percentage of studies with poor or moderate image quality (score 0 and 1) was significantly lower for TEE (9.0%) compared to SHI (24.3%; p < 0.001 vs. TEE) and FI (45.9%; p < 0.001 vs. SHI). For patients who demonstrated a negative FI study but a positive SHI study, the number of studies with low image quality (score 0 and 1) using the FI modality (n = 13) was significantly higher as compared to the number of studies with low image quality using the SHI modality (n = 4; p < 0.005). No significant difference in image quality existed between TEE and SHI in the studies in which atrial shunting was detected either only by TEE or by SHI.
This study demonstrates in a large cohort of patients evaluated for suspected atrial right to left shunt that TEE and SHI had a similar success rate and were clearly superior to FI for the detection of atrial right to left shunt after venous injections of an echocardiographic contrast agent. The TEE missed a few cases with atrial right to left shunt detected only by SHI, and vice versa, indicating that the two imaging modalities should be considered as complementary to each other for the detection of atrial shunting.
TEE and TTE for the detection of atrial right to left shunt
Patent foramen ovale is a frequent finding in the normal population. In autopsy series, the prevalence of PFO is reported to range from 25% to 34% (20). Transthoracic CE has been reported to detect a PFO in 10% to 24% of normal persons (1,2,21,22). Transesophageal CE detects atrial shunting through a PFO in about 22% to 39% of cases (23–26). In comparative studies, TTE has been shown to be inferior to TEE for the detection of PFO (8,12,23,26–28), allowing visualization of an atrial shunt in only about 31% to 47% of positive TEE studies (12,23,28). Factors limiting the sensitivity of TTE to detect PFO have been associated mainly with the impaired image quality (12,28). Thus, TEE has been considered as the diagnostic method of choice for the evaluation of atrial shunting (5,8,29,30).
Combining TTE with SHI, a new imaging modality designed for the improved visualization of echocardiographic contrast agents, we found TTE and TEE to have a similar yield for the detection of atrial right to left shunt. Atrial shunting was not detected by SHI in five patients with a PFO who clearly showed an atrial right to left shunt by TEE. Conversely, we could demonstrate that atrial shunting was not detected by TEE in six patients with a PFO who clearly demonstrated contrast passage using SHI. This is a new finding, for in previous reports the TEE studies were positive in 100% of patients with a positive TTE study (13,23). Two possible explanations may account for the finding of a negative TEE study in a patient with a positive TTE study. First, there is a true existing PFO that is not detected by TEE (false negativeTEE study). The most likely explanation is failure to perform the Valsalva maneuver strenuously during the TEE study, as has already been suggested by others (14,15). Sedation and esophageal intubation may be important limiting factors to perform a strenuous Valsalva maneuver. Second, there is no existing PFO (true negativeTEE study) but contrast bubbles can be visualized by TTE. In this case the presence of pulmonary shunts may lead to a delayed (>3 cardiac cycles) appearance of smaller bubbles in the left atrium (31). This may result in the erroneous misinterpretation of a TTE study as being positive.
In our study, patients with a negative TEE but a positive SHI study showed evidence of atrial right to left shunt within the first three beats after contrast appearance in the right atrium either spontaneously or after the Valsalva maneuver. Because of the close time sequence of contrast appearance in the right and left atrium, pulmonary shunting is excluded in these patients. Therefore, the TEE studies were classified as being false negative. It should be noted that in patients with a negative TEE study but a positive SHI study, shunt magnitude by SHI was rather small, indicating that the clinical relevance of such a finding may be questionable (32).
Impact of SHI on the detection of atrial right to left shunt
Second harmonic imaging is a modality that has been originally developed to enhance the visualization of transpulmonary contrast agents in the left heart chambers using the nonlinear properties of contrast microbubbles. It is based on the principle of receiving double the emitted ultrasound frequency, forming an image from the second harmonic component of the backscattered signal. More recently it has been shown that even without the use of contrast agents, SHI significantly improves the definition of endocardial borders in difficult-to-image patients compared with the fundamental mode (16,17). Although the mechanism is still not fully understood, improvement in image quality with noncontrast harmonic imaging has been associated with 1) reduction in near-field clutter, which is almost entirely made up of ultrasound energy at the fundamental frequency, 2) reduction in the side lobe level and main lobe width (33,34), and 3) higher receiver frequency improving overall spatial resolution. The most apparent effect of SHI is an increase in the signal-to-noise ratio enhancing the visualization of tissue. In our study the use of SHI resulted in a significant improvement of image quality compared with FI, leading to a marked reduction of studies with poor or moderate image quality. Improvements in image quality was not only seen in the near field but also in the far field of the imaging sector, which led to an improved delineation of the anatomy and pathomorphology of atrial structures.
The visualization of microbubbles was clearly enhanced using SHI, resulting in a brighter, larger and stronger backscattered ultrasound signal. This helped in the identification of contrast bubbles even in those patients in whom image quality were not optimal. Although we did not systematically analyze the effect of SHI on our contrast agent, namely an agitated air-filled polygelatin solution, the mechanism for this finding should be bubble resonance or bubble destruction similar to the effect of SHI on other air-filled contrast agents. Moreover, owing to the fact that these bubbles were not capable of passing through the pulmonary capillary bed, microbubble size is relatively large with this contrast agent. Because the backscattering properties of microbubbles are related to the sixth power of the radius, a strong signal can be expected when these large bubbles are shunted into the left atrium. This effect of harmonic imaging on bubble visualization may explain the larger shunt volume detected by SHI compared to FI, the larger duration of contrast visibility compared to FI and TEE and the more frequent visualization of contrast compared to FI.
Impact of image quality on the visualization of atrial right to left shunt
In our study we found a significant relationship between the detection of atrial shunting and image quality by FI. In patients with a negative FI study but a positive SHI study, the number of studies with low image quality (grades 0 and 1) using the FI modality was more than threefold larger compared to the number of studies with low image quality using the SHI modality (13 vs. 4; p < 0.005). This finding may be explained by the increased background noise and clutter with impaired image quality in the FI images in which the microbubbles may have been easily missed. Image quality with TEE was significantly higher compared with SHI. However, the crucial advantage of SHI is its unique impact on the visualization of contrast agents, which should therefore be considered as the ideal imaging modality for the detection of echocardiographic contrast agents. This reduces the advantage of TEE using contrast agents for the detection of atrial shunting and may explain the comparable yield of SHI and TEE for the detection of atrial right to left shunt.
Significance of the study
Transesophageal echocardiography has become the method of choice for the evaluation of patients with otherwise unexplained stroke owing to its increased sensitivity to detect possible cardiac sources of emboli including left atrial thrombi, vegetations, PFO, atrial septal aneurysm, spontaneous echo contrast, mitral valve disease, and significant atherosclerotic disease of the aorta (12,23,26,29,35). The high incidence of cerebrovascular events has prompted the evaluation of possible cardiac sources of emboli as one of the most frequent indication for TEE studies. However, in young patients without clinical evidence of cardiovascular disease, the spectrum of possible cardiac sources of emboli is confined to PFO, atrial septal aneurysm, spontaneous echocardiographic, and degenerative changes of the mitral valve such as myxomatous degeneration, valve prolapse and valvular strands (8,26,36,37). For PFO a causative relationship to ischemic neurologic events has been suggested (3–5). Although therapeutic measures have been advocated (38–41), in clinical practice anticoagulation or closure of the PFO should probably be reserved to patients at highest risk of recurrences. The clinical relevance of an isolated atrial septal aneurysm, spontaneous echocardiographic or degenerative mitral valve disease is less well established, and whether therapeutic consequences should be drawn from these echocardiographic findings is still an unresolved question.
Although TEE has been reported to detect possible cardiac sources of emboli in as much as 39% of patients without clinical cardiac disease, intracardiac thrombi or spontaneous echocardiographic, is very infrequent in these patients (12,42). In a previous study it has been speculated that one-third of TEE studies could have been eliminated had patients without atrial fibrillation or history of cardiac disease been excluded from the study population (26). In a large study with more than 800 patients, TEE was reported to have a low yield for left atrial thrombus, spontaneous echocardiographic or complex aortic atheroma in patients with a normal TTE study and sinus rhythm (42). Thus, in young patients without clinical evidence of cardiac disease or atrial fibrillation, the additional value of a TEE study as compared to a TTE study is mainly confined to the detection of an atrial right to left shunt associated with a PFO or atrial septal defect (ASD). Therefore, Leung et al. (42)recommended a TEE study in patients with an abnormal TTE study and in younger patients when the finding of a PFO contributes to patient management. Because TEE and TTE using SHI have a similar yield to detect atrial shunting, the use of SHI may obviate the need to perform a TEE study in these patients.
Several methodological limitations of the study have to be considered. A complete blinding of the TTE studies was not possible because SHI images can be identified as such by an experienced observer. However, the investigators were not aware of the patients’ identity during data analysis. Moreover, because of the random order of the TTE studies on the videotape a direct comparison between FI and SHI studies of the same patient was made impossible. The sometimes blooming appearance of the microbubbles with SHI made it difficult to differentiate single bubbles from a cluster of adjacent bubbles in stop-frame images, especially in studies with more intensive contrast filling. Similarly, with FI it was sometimes difficult to differentiate the smaller and tinier-appearing microbubbles from the underlying clutter speckles in stop-frame images. We sought to keep the patient body position constant during the different contrast injections to avoid bias by change of body posture. In the left lateral decubitus position, however, right to left shunting may be impaired because of the unfavorable pressure gradient forcing contrast bubbles in the overlying right atrium to cross the atrial septum against an elevated pressure gradient into the left atrium (43). The lack of a true in vivo gold standard for the detection of PFO is another limitation. Although considered as a “gold standard” the true diagnostic accuracy of TEE to detect PFO is not known, and previous reports have indicated that CE using TEE may underestimate the true prevalence of atrial right to left shunt (43,44).
Both SHI and TEE have a similar yield in the detection of atrial right to left shunt using CE. Therefore, in young patients with stroke without clinical evidence of cardiac disease or arrhythmia in whom the indication for a TEE study is the identification of PFO, a positive SHI study may obviate the need to perform a TEE study. In this patient subset the noninvasive SHI should be used as the first imaging modality. A TEE study should be added in case of a negative SHI study and ongoing suspicion of other cardiac sources of embolism.
☆ No financial support was received for this study.
- atrial septal defect
- contrast echocardiography
- computed tomography
- fundamental imaging
- left ventricular
- magnetic resonance imaging
- patent foramen ovale
- second harmonic imaging
- transesophageal echocardiography
- transthoracic echocardiography
- Received January 20, 1999.
- Revision received June 14, 1999.
- Accepted August 5, 1999.
- American College of Cardiology
- Cujec B,
- Polasek P,
- Voll C,
- Shuaib A
- Teague S.M,
- Sharma M.K
- Dubourg O,
- Bourdarias J.P,
- Farcot J.C,
- et al.
- Pearson A.C,
- Labovitz A.J,
- Tatineni S,
- Gomez C.R
- Hausmann D,
- Mügge A,
- Daniel W.G
- Schiller N.B,
- Shah P.M,
- Crawford M,
- et al.
- Hanley P.C,
- Tajik A.J,
- Hynes J.K,
- et al.
- Louie E.K,
- Konstadt S.N,
- Rao T.L,
- Scanlon P.J
- Lee R.J,
- Bartzokis T,
- Yeoh T.K,
- et al.
- Di Tullio M,
- Sacco R.L,
- Venketasubramanian N,
- et al.
- Zenker G,
- Erbel R,
- Krämer G,
- et al.
- Tice F.D,
- Slivka A.P,
- Walz E.T,
- et al.
- Bridges N.D,
- Hellenbrand W,
- Latson L,
- et al.
- Leung D.Y,
- Black I.W,
- Cranney G.B,
- et al.
- Siostrzonek P,
- Lang W,
- Zangeneh M,
- et al.