CRITICAL APPRAISAL OF TRANSESOPHAGEAL ECHOCARDIOGRAPHY

REVIEW

Critical Appraisal of Transesophageal Echocardiography: Limitations, Pitfalls, and Complications James B. Seward, MD, Bijoy K. Khandheria, MD, Jae K. Oh, MD, William K. Freeman, MD, and A. Jamil Tajik, MD, Rochester, Minnesota

Because transesophageal echocardiography is invasive, it has the potential for serious complications. Limitations occur because of the restricted transducer mobility within the confines of the esophagus and stomach and the inability to easily exchange transducers for special situations. Pitfalls (potential erroneous diagnoses resulting from misinterpretation of normal and abnormal anatomy) are prevelent with this new technology. This report critically reviews transesophageal echocardiography and discusses and illustrates commonly encountered limitations, pitfalls, and complications. (JAM Soc EcHOCARDIOGR 1992;5:288-305.)

T ransesophageal echocardiography (TEE) has be-

MATERIALS AND PATIENTS

diographic laboratory. This number represents 5% of all standard transthoracic echocardiographic examinations performed at our institution. The population examined ranged in age from 9 to 92 years with the mean age of 62 years; 2152 were men (56%). The TEE technique and anatomic correlations have been previously reported. 1•2 Clinical indications for TEE are shown in Figure 1. Prospective documentation of each procedure with special attention to complications was tabulated throughout this period. The results of these data are discussed relative to type and incidence of major and minor complications. Pitfalls, which deal primarily with misinterpretation of new and unique observations, were also noted in this same population. However, pitfalls once recognized and categorized became commonplace observations. Thus the incidence of specific pitfalls is not known but is presented to inform the training physician of potential errors or confusion in TEE interpretation. Limitations of TEE are similarly discussed and highlight those portions of the examination that may be incomplete because of unique imaging or physical constraints.

Between November 1987 and October 1991, 3827 TEE examinations were performed in our echocar-

LIMITATIONS

come the first widely used invasive cardiovascular ultrasound examination to be considered a logical extension of a standard echocardiographic study in a select group of patients. 1 •2 Although the quality of the TEE image is consistently superior and can be obtained in the great majority of patients, there are limitations to this technique because of the limited window within the confines of the esophagus and stomach and the inability to easily exchange transducers in special situations. Pitfalls, potential erroneous diagnoses resulting from misinterpretation of normal and abnormal anatomy, are prevalent because of the new tomographic presentation and necessity of using transducers not optimized for all diagnostic situations. Last, because TEE is an invasive examination, there is a small but real incidence of minor and major complications. 3-6 This report critically reviews the current practice of transesophageal echocardiography with particular emphasis on limitations, pitfalls, and complications of this procedure.

From the Division of Cardiovascular Diseases and Internal Medicine, The Mayo Clinic. Reprint requests: James B. Seward, MD, Mayo Clinic, 200 First Street, SW, Rochester, MN 55905. 27/l/36664

288

General

TEE has introduced a new era of clinical invasive ultrasound?· 8 The logistics and concerns of an invasive procedure are offset by consistently high-qual-

Volume 5 Number 3 May-June 1992

Critical appraisal of TEE

Study/ Mise (2%) protocol (4%) Critically ill (3%) Aortic path. (6%)

289

Infarction/ ischemia (3%) Source of embolism (32%)

CHD (8%)

Tumor mass (8%) Prosthetic valve (8%)

----Native valve· disease (13%) Endocarditis (13%)

Figure 1

TEE indications in 3827 procedures. CHD, Congenital heart disease.

ity diagnostic images. However, in young and elderly patients, TEE has unique considerations and potential limitations. Children under the age of 10 years usually require general anesthesia to undergo the examination. 9 - 11 Elderly patients are much more sensitive to systemic sedation and have a higher incidence of esophageal and cervical neck pathologic conditions that need to be specifically addressed with each examination. 12 In our series of 382 7 consecutive examinations, the procedure was aborted in 42 patients ( 1.1%) because of unsuccessful intubation in 36 patients (0.94% ), patient intolerance (5 patients), and esophageal pathologic condition ( 1 patient). Imaging Limitations New tomographic views. TEE is often referred to as a new window to the heart. As such, the tomographic anatomy visualized is new and often initially confusing. To the beginner, the examination is unfamiliar and consequently may pose limitations regarding interpretation, demonstration of appropriate anatomic relationships, and diagnostic application. It is important to relate this new echocardiographic examination to existing experience. A level II echocardiographer is very familiar with two-dimensional tomographic echocardiographic anatomy as obtained from a surface examination. The TEE examination obtains the identical images once the orientation techniques are mastered. The transducer is posterior to the heart, thus we recommend orienting all images such that the transducer burst is at the bottom of the viewing screen. This orients the short-axis, long-axis, and apex-up four-chamber views identical to those

obtained from the precordial examination. A uniform presentation of tomographic views regardless of transducer position or electronic bang fosters better understanding and reduces inherent limitations and pitfalls, which will be discussed. If we abandon established conventions and rules, we expose ourselves to multiple problems of image presentation, structure recognition-identification, and anatomic correlations. Images of familiar structures become difficult to intuitively recognize if the images are presented in an unfamiliar format (e.g., aortic valve and left ventricular short axis as mirror image upside-down presentations). Unconventional image orientations have produced and will continue to lead to potentially significant errors of interpretation. There is no logic in changing the entire image orientation just because the image transducer is in the esophagus rather than the chest wall. Imaging planes that one obtains by TEE and transthoracic echocardiography are identical (long-axis plane, short-axis plane, and four-chamber plane). A series of views of the heart can be obtained along any of these primary planes by TEE. Therefore, it behooves all of us to uphold the American Society of Echocardiography standard of image orientation for the long-axis, short-axis, and four-chamber views irrespective of the position of the transducer or electronic bang. If we as echocardiographers have difficulty with our "own images," how can we effectively communicate with our colleagues less familiar with echocardiography. Air. The walls of the esophagus and stomach are the portal through which the heart and great vessels

290

Journal of the American Sociery of Echocardiography

Seward et al.

of the thorax are visualized during the transesophageal examination. Although images are consistently superior to comparable surface images, there is limited maneuverability within the lumen of the esophagus and fundus of the stomach. Entrapped air consistently interferes with certain transducer maneuvers, particularly retroflexion of the scope tip in the midesophagus and imaging from the fundus of the stomach. Entrapped air rarely precludes a diagnostic examination. Biplanar and multiplanar13 transducers compensate for this limitation; however, in any particular examination all areas of the heart may not be imaged with consistent predictability or quality. There is no complete solution to the interference of entrapped air; however, there are some helpful observations. First, there is more air in the esophagus in the beginning of the examination than at the end because of continual clearing of air by esophageal and transducer action during the study. Early in the procedure it is best to use views and scope manipulations that are not as likely to be affected by entrapped air. Anteflexion, movement of the transducer lens into apposition to the esophageal wall, is preferred over retroflexion, which moves the transducer away from the esophageal wall. Retroflexion more easily permits entrapped air to come in front of the transducer lens. Early in the examination, try the longitudinal transducer or short-axis views with the transverse plane, which uses anteflexion for best images. Obtain four-chamber views, which require retroflexion, later in the examination. Passing the scope tip into the stomach also tends to clear air from the esophagus. Additionally, in the same sized patient, larger scopes have greater esophageal contact and less air interference than smaller scope heads. Air-filled bronchus and trachea. The trachea and air-filled bronchi consistently interfere with certain tomographic planes of section. Basal cardiac and aortic arch examinations are particularly affected by the interposed air-filled trachea and bronchi. This problem can be largely circumvented with multiplanar transducers, which allow off-axis imaging. Attempts to image the upper ascending aorta using the horizontal plane is limited by a "blind spot" caused by the interposed bronchus between the esophagus and upper ascending aorta. 1 This limitation can be minimized by using an anteflexed longitudinal plane that directs the ultrasound beam over the left bronchus.2 Similarly, the distal left pulmonary artery, which can be obscured by the left bronchus, is best imaged in the off-axis longitudinal plane. Equipment

Current TEE equipment consists of a two-dimensional echocardiographic instrument and an endo-

scope fitted with an ultrasound transducer. It is inconvenient to change transducers, alter frequently, or incorporate offset capability. Present transesophageal transducers are optimized only for midfield imaging. An ideal scope or family of scope should be able to image the extremes of near and far field and incorporate multifrequency transducers, offset capability, and other innovations that would reduce these limitations. Newer esophageal scopes will be smaller in diameter, 14 incorporate continuous-wave Doppler, and ultimately have wide-field image 15 •16 capability. Tissue characterization17- 19 and border recognition, as well as three- and four-dimensional scanning, 20•21 remain illusive but potentially very useful innovations. As with any technology that is in a state of evolution, certain limitations are inherent with the "state of the art." Distinct advantages that have driven the clinical application of TEE include ( 1) consistently superior images, and (2) low incidence of unacceptable examinations. Equipment limitations can be divided into those of transducer or scope design. Transducer Limitations Scan plane. The original monoplane 1 (e.g., horizontal plane) is now largely being displaced by hiplanar (e.g., horizontal and longitudinal plane) transducers. 2 Because the transducer is confined to the esophagus and fundus of the stomach, multiplanar22 and ultimately omniplanar 13 scopes will become state of the art. New image presentation and inherent technologic constraints continue to press for more views because of the comparatively limited mobility of the esophageal transducer. Near- and far-field imaging. Most esophageal transducers are 5 MHz and are optimized for visualization of midfield structures such as the atrial septum and mitral valve. Far-field limitations include ( 1) inability to optimally visualize the ventricular apex from the midesophagus, (2) acoustic shadowing of anterior aortic annulus and cardiac crux by the central fiberous body of the cardiac valves, and (3) far-field shadowing caused by intervening prosthetic material (e.g., mitral prosthesis obscuring the left ventricular outflow tract or aortic prosthesis obscuring the anterior aortic root) (Fig. 2). Near-field limitations include (1) transducer artifact interfering with visualization of the lower pulmonary veins, and (2) poor visualization of the pediatric heart because of inappropriate transducer frequency and focal point. Multifrequency transducers, offset capability, and specialty scopes will ultimately reduce these limitations. Image. Miniaturization or incorporation of mul-

Volume 5 Number 3 May-June 1992

Critical appraisal of TEE 291

tiple scan planes often requires a reduced number of imaging crystals per scan plane, which degrades the quality of the resultant image. New transducers are being developed that optimize the relationship of transducer size and sophistication with greatest image quality. These advances are accomplished by multifrequency transducers, increased numbers of imaging crystals, and incorporation of multidirectional chains. Thus limitations in the image quality will be directly related to equipment state-of-the-art. Scope Limitations General. The delivery device for TEE is a modified endoscope, fitted with an ultrasound transducer. The scope length, depth markings, control knobs, lock mechanism, and handle design are, in most instances, not optimized for cardiologic applications. The shape of the transducer, obligatory electronics, and physiologic monitoring capability are unique to TEE. Depending on the application, any of the above can be a potential limitation. Devices specifically suited for TEE application will reduce the annoying constraints of current technology. Size. When TEE was first introduced, an adultsized (9-mm shaft) endoscope tipped with an ultrasound transducer was used. With increasing sophistication of the technique, many scope shafts have actually increased in size ( ± ll-mm diameter shaft). (Note: An increase in shaft diameter from 9 mm to ll mm increases the cross-sectional area of the shaft by 48%, whereas a decrease in shaft diameter from 9 mm to 7 mm reduces the cross-sectional area by 40%.) Additionally, the transducer tip usually has an even larger cross-sectional area. To our knowledge, no specific complications have been directly attributable to the size of the scope. 23•24 However, transient vocal cord paralysis has been reported in two patients after lengthy neurosurgical procedures in the upright position that required extreme anteflexion of the patient's head with simultaneous use of a large (9-mm shaft) transesophageal scope alongside an endotracheal tube. 25 This complication has not recurred with use of smaller diameter scopes. We have also experienced an increased incidence of inability to introduce larger scopes into some elderly patients, small adult or pediatric patients, and intubated patients. Extreme flexion of the transducer tip potentially can injure or perforate the esophagus. 26-28 Smaller scopes are now commercially available. However, smaller scopes presently incorporate fewer imaging crystals and have fewer controls and reduced planes of imaging. Concerns for transducer contact and a higher incidence of complications with smaller scopes have not been substantiated to date. Maintenance. A number of poorly documented

Figure 2 Shadows. Far-field structures can be shadowed by interposed ultrasound reflectors. Prosthetic material, calcium, and fibrous cardiac skeleton are common reflectors interfering with far-field image. In this figure, a mitral bioprosthesis (P) and sewing ring (white arraws) produce linear shadows in the far field (black arrows). Note the upper ventricular septum (VS), free wall of right ventricle (R V), and apex of left ventricle (LV) are shadowed. LA, Left atrium.

limitations have been encountered with scope maintenance. The cables that control the movement of the tip frequently stretch with time. Early devices had an unacceptably high incidence of excessive stretch and fracture of cable connections. These limitations have been substantially reduced. However, periodic readjustment of cable tension may still be necessary even with ideal maintenance. PITFALLS False Masses TEE images structures heretofore inaccessible by transthoracic echocardiography. Thus unfamiliar normal anatomy may be misinterpreted as abnormal (Figures 3, 4, and 5). Trabeculations of the atria or atrial appendages. Muscular trabeculations (i.e., pectinate muscles) and irregularities are seen in the walls of both atrial appendages, and the examiner should become well acquainted with their normal appearances (Figure 6). These muscle ridges are usually small and refractile, move in concert with the atrial wall, and are typically multiple. However, thrombus is char-

292

Seward et a!.

Figure 3 Spinal cord. By rotating transducer posteriorly, spinal cord (SC) can be visualized at each neuro-interspace of the thoracic spine. Spinal fluid (SF), spinal cord (SC), and vascular pedicle can produce a mass-like appearance. This observation can be misinterpreted as an extracardiac mass or tumor.

acteristically of different texture than atrial wall and is more or less echo refractile, uniform in consistency, often pedunculated, and typically .occurs in conjunction with significant atrioventricular valve disease or low output state. Tumors that extend into the atria or atrial appendage are also distinctly different in echo density from normal muscle ridges. We have found that multiplanar imaging is very helpful in recognizing atrial appendage pathologic conditions. Within the left atrium, the orifices of the pulmonary veins are encircled by tissue that may appear masslike in some tomographic planes of section. One of the most frequent mass effects is encountered is the common wall separating the left atrial appendage from the left upper pulmonary vein (Figure 7) . In a tomographic plane of section, the terminal portion of this partition appears globular and looks like a mass, especially as it undulates with cardiac motion. This common wall in its midportion is thin, and the globular end can sometimes be quite large, mimicking a left atrial tumor. Awareness of this common anatomic variant should obviate serious misinterpretation. A persistent left superior vena cava 29 and a potentially fluid-filled recess of the pericardia! reflection course within this common wall and should not be misinterpreted as a cyst, abscess, or other abnormal structure (Figure 5, a and c). Notably, a persistent left superior vena cava is associated with a large coronary sinus that normally enters the right

Journal of the American Society of Echocardiography

Figure 4 Pseudomass. This transgastric long-axis view of left ventricle (LV) is from patient with mitral prosthesis. Mobile transected chordae within the left ventricle (arrows) give a pedunculated mass effect. These transected chordae appear multicentric, mobile, bulbous, and pedunculated, features also associated with thrombus. However, there were no significant wall motion abnormalities, and the true identity as chordae could be discerned by long- and shortaxis transgastric images. PW, Posterior wall; A W , anterior walL

atrium 1•29 and has color-flow detectable blood movement within the space. Pericardia! fluid within the pericardia! reflection will not generate a color-flow signal. In the right atrium at the orifices of the superior and inferior vena cava, muscle bundles can also appear mass-like (Figure 6) . In the horizontal plane at the orifice of the right superior vena cava, an encircling muscle ridge will appear ovoid and masslike. Slow withdrawal of the endoscope to the lumen of the superior vena cava completes the muscle ridge around the orifice of the superior vena cava and ensures proper identity. Less commonly, a similar ridge of muscle is visualized at the orifice of the inferior vena cava. In the horiwntal plane, advancing the scope will complete the encircling eustachian valve around the oriface of the inferior vena cava. Biplane or multiplane imaging consistently eliminates confusion. Atrial septum. The atrial septum surrounding the · centrally located membrane of the fossa ovalis is fat laden and in older patients normally can be up to l em thick (Figure 8). In certain tomographic planes, the limbus of the fossa ovalis can appear masslike. When excessively thick, it is referred to as lipomatous

Volume 5 Number 3 May-June 1992

Critical appraisal of TEE 293

Figure 5A Normal cyst-like structures-transverse sinus (TS). Top panel) Long-axis longitudinal scan of proximal ascending aorta (AscAo). The TS surrounded anteriorly by theAscAo) superiorly by the right pulmonary artery (RPA) J and posteriorly by the left atrium (LA). Bottom panel) Shortaxis longitudinal scan at level of the aortic valve (A V). The TS is bounded posteriorly by the LA) medially by the A VJ and anteriorly and internally by the main pulmonary artery (MPA). LJ left coronary cusp; RJ right coronary cusp; N, noncoronary cusp, RAJ right atrium; R VOJ right ventricular outflow tract; L VOJ left ventricular outflow; VSJ ventricular septum.

Figure 5B Membrane of the fossa ovalis. Top panel) Horizontal scan at level of the proximal ascending aorta (AscAo). Membrane of fossa ovalis (white arrows) overlaps posterosuperior limbus of atrial septum producing an interposed cavity. When the foramen ovale is patent, blood can cross between right atrium (RA) and left atrium (LA) through this potential orifice (valve of the fossa ovalis). Bottom panel) Longitudinal scan at membrane of fossa ovalis (white arrowheads). This membrane overlaps the superior margin of atrial septum I posterior wall of the AscAo (black arrows) producing an interposed space. SVC, Superior vena cava. (All other abbreviations same as SA.)

hypertrophy or atrial septallipoma30 and usually appears as an echodense mass of variable size and consistency, that occasionally may reach large proportion. Particular confusion arises when the lipomatous hypertrophy of the atrial septum is asymmetric. This pitfall is more common with a monoplane TEE examination and is clarified by a biplane examination. This condition is benign but can be easily misinterpreted by an inexperienced examiner. Tricuspid annulus. The tricuspid annular sulcus at the free wall of the right ventricle is filled with variable amounts of fatty tissue (Figure 9). This tissue

can produce a mass effect in the tricuspid atrioventricular groove, particularly when viewed obliquely in the horizontal tomographic plane. This common observation should not be misinterpreted as a tumor or ring abscess. Magnetic resonance imaging has also been diagnostic when the observation remains in doubt. Mitral valve annulus. The mitral valve annulus is usually less problematic; however, annular calcification and occasionally fat can be very impressive and appear masslike. Because of the near-field location, occasionally calcium may not be as dense appearing

294 Seward et al.

Figure 5C Top panel) Right atrial appendage (RAA). In long-axis longitudinal scan of ascending aorta (AscAo) the RAA can be imaged anterior to AscAo and should not be n_llstaken for a p~thologic observation. Bottom panel) Persistent left supenor vena cava (LSVC). Partition between left upper pulmonary vein (LUPV) and left atrial append~ge (LAA) can_ appear as a space. Two circumstances typKally cause th1s appearance. (l) A pericardia! reflection within this. Part:ition b~comes fluid-filled in the presence of pencardial effus10n. Usually this is identified as fluid and associations such as fluid-filled transverse sinus and pericardia! sa~. ~2) Persistent LSVC (example shown) that courses w1thm the common wall is identifi_ed by ~olor-flow Doppler blood flow and usually assoClated dila~ed coronary sin~s. 29 LUPVJ Left upper pulmonary vern; LV, left ventncle. (All other abbreviations same as SB.)

and may be mistaken as a tumor. However, the characteristic location and reflectance of calcium usually allows easy recognition. Aortic valve. A frequently observed mass effect occurs when the aortic valve is cut obliquely in the short axis. The ovoid apperance of the aortic sinus can be mistaken for vegetation or tumor. Similarly,

Journal of the American Society of Echocardiography

Figure 6 Right atrial (RA) trabeculations or pectinate muscles (arrowsheads) and muscular bands (arrows). The right atrium is more trabeculated (i.e., pectinate muscle~) than the left atrium (LA). A highly trabeculated atr1al appendage may be d~ffirult to clearly distinguish from thrombus. Example 1s from patient with atrial se~tal defect in whom the atrial musculature is hypertrophied. Top panel) A larger muscle band (arrows) is commonly visualized at orifice of the superior vena cava. Bottom panel) The identity of normal atrial muscle bundle can be obtain~d by slo~ withdrawal of the transesophageal scope m the honwntal plane. Normal muscle encircles the s_uperior_ vena caval orifice (SVC)J separating it and the nght atr1al appendage. AS) Atrial septum; A VJ aortic valve.

the aortic valve can be incorrectly interpreted as either bicuspid or tricuspid when cut obliquely. The pitfalls occur when using a single imaging plane; biplane and multiplane views of the aortic valve eliminate this pitfall.

Volume 5 Number 3 May-June 1992

Figure 7 Pseudomass. Normal partition between left upper pulmonary vein (LUPV) and left atrial appendage (LAA) can be fat laden and appear masslike (laJEe arrows). Proximal portion of common wall is usually thin (small arrows) and distal portion bulbous (laJEe arrows), adding to mass effect. AscAo, Ascending aorta; PA, pulmonary artery, LA, left atrium.

Other structures. Surgical sutures can be visualized at the sewing ring of prosthetic valves or at the margins of prosthetic patch material (Figure l 0). Redundant sutures can appear filamentous or pedunculated and undulate with the cardiac motion. The regular spacing and highly refractile appearance of sutures help make a correct interpretation. Although redundant suture material will move with the cardiac cycle, these structures should not appear multicentric, irregularly spaced, excessively mobile, elongated, or bulbous (features more consistent with adherent thrombus or vegetation). Other intracardiac structures, which in a tomographic plane look like a mass, include pacing wires, retained catheters, prosthetic patch material, and artificial valves and their sewing rings. Membranes Membrane of the fossa ovalis. This membranous portion of the atrial septum can be redundant, aneurysmal, and show variable undulating motion with each cardiac and respiratory cycle. 31 Occasionally, if the membrane has large excursion, it may produce a mass effect in the left atrium, particularly with a monoplane examination. Biplane imaging easily elucidates the true identity of this structure. Valve of the fossa ovalis. The posterosuperior margin of the fossa membrane overlaps the superior

Critical appraisal of TEE 295

Figure 8 Lipomatous atrial septum. The atrial septum frequently becomes inundated with fatty tissue (i.e., lipomatous hypertrophy. 30 A characteristic mass effect is observed. Membrane of fossa ovalis (small arrows) is spared, while fatty atrial septum (AS) is thickened with hyperechoic fat (laJEe arrowheads). A pathopneumonic dumbbell shape of the atrial septum is observed. Amount of fatty infiltrate is variable but can be impressive. This condition is usually considered benign. LA, Left atrium; LV, left ventricle; RA, right atrium; R V, right ventricle; VS, ventricular septum.

fatty limbus of the atrial septum (i.e., the valve of the fossa ovalis). N onfusion of these structures results in a patent valve of the fossa ovalis. The overlap between the fatty atrial septum and the fossal membrane in certain tomographic planes (particularly the horizontal plane) appears as a cavity29 (Figure 5, B). If the valve of the fossa ovalis is patent, shunting from one atrium to the other can be observed within this space. Longitudinal planar images best delineate this potentially confusing anatomy. Left atrial membrane. A thin membrane or partial form of cor triatriatum extends from the common wall separating the left upper pulmonary vein and left atrial appendage and crosses the atrial cavity to the superior limbus of the foramen ovale29•32•33 (Figure ll). This membrane is variable in its extent but is usually incomplete and nonrestrictive. Eustachian valve. In the right atrium at the orifice of the inferior vena cava, the eustachian valve often appears as a mobile undulating membrane or mass partially encircling the orifice of the inferior vena cava as it enters the floor of right atrium33 (Figure 12). Accurate delineation is best appreciated with biplane or multiplane imaging.

296

Journal of the American Society of Echocardiography

Seward et al.

Figure 9 Pseudomass of the tricuspid valve (TV) annulus (arruws) . Tricuspid annulus at the right ventricular (R V)right atrial (RA) junction normally contains fat that can give a mass effect (arruws) (particularly noticeable in the horizontal image plane). VS, Ventricular septum; LV, left ventricle; AS, atrial septum; MV, mitral valve, LA, left atrium.

Echo-free Spaces Transverse sinus. The transverse sinus, a pericardia! reflection between the left atrium and the great vessels at the base of the heart, typically contains a small amount of pericardia! fluid that produces a small, crescent-like, echo-free space between the left atrium and aorta, as visualized in the horizontal plane, and a triangular space in the longitudinal planes 2 (Figure 5, A). When filled with larger amounts of fluid, this space can be misinterpreted as a pathologic finding and confused with a cyst or abscess cavity. Within the fluid-filled space, the left atrial appendage, its outpouchings, and attached epicardial fat can appear as mobile cystic or solid masses (Fig. 13). Biplanar imaging and slow sweeps of the ultrasound across the sinus and atrial appendages usually permits appropriate identification. Oblique sinus. The oblique sinus, a posterior pericardia! reflection between the pulmonary veins, can present as a fluid-filled space interposed between the left atrium and esophagus. A pericardia! cyst can also appear in the same position. Recognition of the pericardia! layers will usually permit appropriate identification of the pericardia! cyst versus fluid-filled oblique sinus. Other pericardial reflections. Because of the new tomographic presentation of anatomy, pericardia! reflections and recesses are easily visualized and

Figure 10 Sutures. Prosthetic material is usually hyperrefractile and causes far-field shadowing. Sutures pose a special problem because they can appear elongated and multicentric. In a prosthetic sewing ring sutures appear as regular (arrowheads) hyperrefractile structures and can act as a nidus for thrombus and vegetation. A suture usually is not pedunculated and does not undulate or become bulbous or appear as multiple filaments (features more typical of pathologic vegetation or thrombus). Arrow, Prothesis; B, ball-(Starr-Edwards aortic prosthesis); LV, left ventricle; LA, left atrium.

are potentially confusing. A pericardia! reflection between the wall separating the left atrial appendage and left upper pulmonary vein can also appear as a cystic space or mass. Hiatal hernia. A large hiatal hernia can markedly interfere with a complete TEE examination34 (Figure 14). However, occasionally a large fluid- or gas-filled hernia will become interposed between the heart and the esophageal lumen. When fluid-filled, the hiatal hernia can appear as a thick-walled cystic mass posterior to the left atrium. When filled with gas, the expected problem with ultrasound propagation and shadowing of more anterior structures makes the transesophageal examination technically difficult. Color-flow Doppler

Because of the close proximity of cardiac chambers and use of higher frequency transducers, color-flow Doppler signals are more exquisitely visualized and alias at low velocities. The increased sensitivity makes the color signals appear more extensive and thus cannot be directly compared with a surface color Doppler examination. However, with experience, reliable semiquantitation of the severity of valvular regurgitation can be obtained, making TEE superior to the

Volume 5 Number 3 May-June 1992

Critical appraisal of TEE

Figure 11 Left atrial membrane. A forme fruste of cor triatriatum appears as an incomplete membrane in the left atrial (LA) cavity. Left panel, four-chamber view: Medially at the atrial septum (AS) the membrane originates from the posterior margin (la'lfe arrows) of the membrane of the fossa ovalis (small arrows). Right panel, basal view at aortic valve (Ao), horizontal plane. Laterally, the cor triatriatum membrane (arrows) inserts onto the common wall (arrowheads) separating the left upper pulmonary vein (LUPV) and left atrial appendage (LAA). LV, Left ventricle; MV, mitral valve; RA, right atrium; R V, right ventricle; VS, ventricular septum; TV, tricuspid valve.

Figure 12 Eustachian valve. Left panel, In the horizontal plane with transducer at the gastroesophageal junction, orifice of the inferior vena cava (IVC) can be visualized. Surrounding the anterior orifice and interposed between IVC and body of right atrium (RA) is eustachian valve (arrows). This structure is usually membranous and undulates with cardiac action. Right panel, In longitudinal plane the eustachian membrane (arrows) originates from the anterior lip of the IVC orifice and separates the body of RA and atrial septum (AS). This undulating membrane can be mistaken for mass or thrombus. A V, Aortic valve; SVC, superior vena cava; LA, left atrium.

297

298

Journal of the American Society of Echocardiography

Seward et al.

Figure 13 Pseudomass in transverse sinus (TS). Within the pericardia! reflection (e.g., transverse sinus) between left atrium (LA) and ascending aorta (AscAo) and pulmonary artery (PA), there is epicardial fat (F, arrows). When TS contains fluid, epicardial fat or the atrial appendage can appear mass-like.

transthoracic examination in most patients. Experience and knowledge of the technology are the most important prerequisites for proper interpretation. Timing of rapid diastolic and systolic events is best accomplished with color-flow Doppler-derived Mmode display or by pulsed-wave Doppler. Quantitative techniques have not received sufficient validation. Regurgitant jets. Systolic flow reversal into pulmonary veins, extent of flow disturbance in the receiving chamber, size of the regurgitant orifice, and hemodynamic and anatomic associations usually allow accurate semiquantitation of color flow jets. 35•36 · Mitral and tricuspid valve regurgitation are best imaged with biplanar examinations, whereas aortic and pulmonic valve regurgitation are best imaged with the longitudinal plane. With the increased sensitivity of TEE, trivial "physiologic" regurgitation, particularly of the mitral and tricuspid valve, is frequently recognized. 1•37 Similarly, the "closing volume" of mechanical prostheses is uniformly observed and must be recognized as a normal observation. 37 The "normal" color flow jets seen in native valves and mechanical prostheses are small and of brief duration. These jets usually represent trivial regurgitant events and do not in themselves prompt endocarditis prophylaxis.

Figure 14 Hiatal hernia. Stomach or bowel can herniate into thorax and become interposed between normal esophagus and heart. Rugal folds (R) within stomach can be mistaken for tumor or mass. Heart can be obscured from visualization, as in this example.

Off-axis regurgitant jets. Two-dimensional echocardiographic color flow imaging is a tomographic examination that allows imaging of blood flow. Because of limited transducer mobility within the confines of the esophagus, regurgitant jets may be imaged incompletely and appear spuriously eccentric or atypical. Misinterpretation of the source or amount of regurgitation can occur with limited tomographic views. It is important to visually confirm the source of any regurgitant jet. The regurgitant source is appreciated as flow convergence and color aliasing of the color signal. Multiplanar imaging, slow sweeping of the color flow image, and confident identification of the source will eliminate the potential misinterpretations. Atrial septal defects. Atrial septal defects are more confidently diagnosed by TEE using color flow imaging. 29 The degree of shunting must be adjusted to the increased sensitivity and not equated with the precordial examination. Accurate measurement of the defect size and recognition of right-sided volume overload are equally important clinically relevant observations. Phenomena Spontaneous contrast effect. Because of the proximity of the transducer to the atrial cavities and the

Volume 5 Number 3 May-June 1992

use of higher frequency transducers, visualization of blood movement within the cardiac chambers is now more commonly observed38•39 (Figure 15). The phenomenon is associated with reduced blood flow and poor clearance of blood from the cardiac chamber, and it is reported that a high incidence of systemic embolism occurs when this phenomenon is observed. This phenomenon, commonly ascribed to rouleau formation, is equated with increased coagulation and stagnation of blood, which can actually look like a thrombus. Reverberations and ghosting. The esophagus is surrounded by air-filled lung. Reverberation signals or ghost artifacts are a common occurrence because of the strong ultrasonic reflection from impedance mismatch between tissue and air. Linear artifacts most frequently occur within the upper ascending aorta and middescending thoracic aorta, which have air-filled lung in the immediate far field (Figure 16). TEE examination of the thoracic aorta is highly diagnostic40·41 Linear artifacts often lie in nonanatomic planes, cross normal anatomy, have artifactual motion, and do not alter Doppler-depicted blood flow or disappear with change in signal depth. Near-perfect duplication of Doppler signals and structures can be obtained in certain imaging planes and is consistently obtained when imaging the descending thoracic aorta (Figure 17). Duplication of the aorta and Doppler signal is obtained if the field of view is expanded to accommodate a second signal. The only way to avoid misinterpretation is to be aware of these phenomenon and to be cautious not to misinterpret atypical anatomy, such as a double thoracic aorta or Doppler signal lying outside of the blood-filled aorta or cardiac chambers. Extracardiac fluid. Accumulations of pleural fluid are easily visualized by TEE. However, loculated fluid may appear ovoid or suspiciously like a normal or abnormal structure (Figure 18). Fibrous bands within the fluid-filled space may give the false appearance of the thoracic aorta with dissection or rupture.42 It is imperative to track all fluid-filled spaces to their source (for example, a ventricular aneurysm to the ventricle or aortic dissection to the normal thoracic aorta). Biplanar and multiplanar imaging usually permit proper diagnosis. Correct identification may be compounded by color-flow artifacts or ghosting phenomenon within the fluid-filled space. These artifacts are usually of low velocity and should not be confused with pathologic blood movement within the fluid-filled space. Because of the proximity of structuresand use of higher frequency transducers, Doppler artifacts are frequently observed within any

Critical appraisal of TEE

299

Figure 15 Spontaneous contrast effect. Top panel, Patient with mitral stenosis and spontaneous contrast effect in the left atrium (LA). Mitral valve (MV) is thickened and the LA enlarged. Spontaneous contrast (slow-moving blood) is visualized within the LA. Bottom panel, More subde contrast effect is noted in the left atrial appendage (LAA) of another patient with less severe mitral valve disease. Spontaneous contrast phenomenon can, in extreme cases, appear as thrombus or mass. An association exists between spontaneous contrast effect and a higher incidence of thromboembolic events. Close observation will distinguish visible nonmovement of clotted blood while the spontaneous contrast effect is distinguished as having slow-moving blood, although the blood products can be visualized. PA, Pulmonary artery; AscAo, ascending aorta; LA, left atrium; A V, aortic valve; LV, left ventricle; R VO, right ventricular outflow; VS, ventricular septum.

fluid-filled space. To avoid misinterpretation, be cautious of interpreting very low velocity signals and always search for a source of communication. Motion artifacts are very common, whereas pathologic com-

Journal of the American Society of Echocardiography

300 Seward et al.

Figure 16 Linear reverberations. Top panel) Horizontal scan and (bottom panel) longitudinal scan of the ascending aorta (AscAo). Frequently in upper AscAo and less commonly in middescending thoracic aorta, linear reverberations (arrows) are observed (thought to occur because of overlying lung). These linear shadows can be confused with pathologic dissecting membranes. Change in transducer frequency, alternating the depth of image, nonanatomical appearance, and nondisturbed color-flow signal can distinguish artifactual nature of this observation in most situations. Occasionally, linear artifacts cannot be adequately distinguished from aortic dissection. RPA) Right pulmonary artery; IAJ left atrium; RAAJ right atrial appendage; SVCJ superior vena cava; .MPAJ main pulmonary artery.

0.2%.43 The incidence of serious complications associated with TEE is reported to be between 0.2% and 0.5%. 3•4 Death is exceedingly rare (0.0098% in 10,218 studies). 3 This rate is comparable to upper endoscopy with complication rates of 0.08% to 0.13% and mortality rate of0.004o/o. 44•45 Both safety and potential complications of TEE have not been extensively reported. 1•46' 53 In our experience, the overall rate of minor and major complications was low at 2.9% (111/3827 procedures) (Figure 19). Complications were grouped into two categories. Major complications (i.e., serious laryngospasm, sustained ventricular tachycardia, congestive heart failure, or death) occurred in 0.2% (9/3827 procedures). There was one death, mortality 1/3827 (0.026%). The remaining complications were considered minor (i.e., brief arrhythmia, transient hypotension or hypertension, minor discomfort, and prolonged sore throat). The patient who died was a 64-year-old, obese, diabetic woman with repeated episodes of congestive heart failure with pulmonary edema. She had a normal exercise thallium study. She was referred for the evaluation of left ventricular systolic and diastolic function as well as assessment of mitral regurgitation. On transthoracic echocardiography, the left ventricular function was normal; however, the evaluation of mitral regurgitation was inadequate, thus TEE was performed. On TEE, the mitral regurgitation was graded as mild to moderate. The patient had received midazolam maleate (Versed) 1.0 mg intravenously as a sedative but was awake with normal pulse oximeter readings. Approximately 10 minutes after withdrawal of the TEE endoscope, the patient suddenly became profoundly dyspneic and hypoxic and ventricular fibrillation developed. Prolonged resuscitation attempts were unsuccessful. Autopsy demonstrated a normal esophagus and normal coronary and pulmonary arteries and lungs. There was evidence of lymphocytic infiltration of the myocardium (mild myocarditis) and mild floppy mitral valve. Irritable myocardium caused by the myocarditis and possible arrhythmia was the explanation for her congestive heart failure. Other Safety Considerations

munication from aneurysm or aortic dissection 1s much less frequently observed. COMPLICATIONS

The risk of complications with upper endoscopy is very low. The rate of serious complications associated with fiberoptic endoscopy ranges between 0.1 o/o and

TEE is unique in many respects. The scope is manipulated predominantly within the esophagus. Resistance to motion may potentiate earlier fatigue of the cables or protective sheath. The electronic imaging system places a burden on the user to also monitor electrical and thermal safety. The scope may carry pathologic organisms or cause disruption of the mucosa resulting in bacteremia. Thennal safety. Ultrasound transducers generate

Volume 5 Number 3 May-June 1992

Critical appraisal of TEE 30 l

Reverberation artifact. A common reverberation artifact occurs when sound beam is directed toward air-filled lung, which acts as an ultrasound reflector. Near-field structures will be duplicated in far-field as a reverberation artifact. All signals are duplicated, including color-flow Doppler. This artifactual duplication of anatomy and Doppler signal should not be mistaken as a true structure. The example is reverberation artifact of descending thoracic aorta (ThAo). In the far field, a reverberation false aorta is imaged when the sound beam is directed toward air-filled lung.

Figure 17

a certain amount of heat. During the standard examination (averaging 20 minutes in the awake patient), transducer heat should not be a problem. Even longer periods of intubation have not resulted in detectable mucosal injury. 24 However, during surgery a scope is left in a single place for extended periods, and the patient is typically cooled. Continuous contact or a significant heat gradient (i.e., patient 28° C and probe 38° C, resulting in a 10° C heat gradient) may increase the chances of thermal injury to the esophagus. Heat-sensing thermisters have been incorporated into all commercial scopes. However, two limitations exist with current thermal sensing circuitry. There is no manual reset, and the range of temperature presets often does not match the clinical situation. Thus extrmely febrile patients occasionally cannot be adequately examined because the device senses the patient's temperature and shuts down. Similarly, intraoperative transducers will shut down on sensing the heat of reprofused blood. An override switch to be used on the judgment of the examining physician would be very helpful in these annoying situations. Thermal gradients (i.e., a significant difference between scope and patient temperature) is not a function of commercially available thermal monitoring. It is our contention that scope thermostats have been more of a hindrance to patient care than the intended function of averting thermal complications. During prolonged monitoring, the ultrasound device should be periodically shut down to allow cool-

Pleural fluid (PF). Fluid accumulation in thorax, particularly around thoracic aorta (ThAo) can be confused as aortic dissection, extravascular mass, or tumor. Proteinaceous filamentous bands (P) can be mistaken for pathologic structures such as tumor or dissecting membrane. Wide-field of view and careful delineation of underlying anatomy and pathology will avoid mistaken diagnoses.

Figure 18

ing. In particular, during intraoperative monitoring, the ultrasound transducer should be inactivated when not being actually used. In febrile patients or situations when an inappropriate autocool signal may oc-

Journal of the American Society of Echocardiography

302 Seward et al.

VT* 1 (0.02%) CHF 2 (0.05%) Death 1 (0.01%)

II 0

Major complication Minor complication

Laryngospasm 5 (0.14%)

NSVT 8 (0.2%)

Transient hypotension 13 (0.3%)

Transient HTN 15 (0.3%)

Hypoxiat 13 (0.3%)

Blood-tinged sputum 9 (0.2%)

* Required defibrillation

t Supplemental 0 2 required

** I.V. verapamil 1 pt

Figure 19 TEE complications in 3827 procedures. CHF, Congestive heart failure; NSVT, nonsustained ventricular tachycardia; SVT, supraventricular tachycardia; VT, ventricular tachycardia.

cur, physicians have put iced saline solution down the esophagus (during operation) or precooled the scope before introduction. More appropriate circuitry logic, comparison of patient and lens temperature, display of scope temperature, and operator override would eliminate some of these limitations. Electrical safety. Perforation or near perforation of the scope skin will result in a loss of "safe electrical grounding." The risk of electrical injury with current devices is extremely remote because the scope is electrically isolated from the ultrasound machine. Thus electrical risk would be a very remote possibility even with a poorly maintained scope. However, an electrical leak may portend a small hole in the scope's skin which may harbor other safety problems. Infectious organisms or caustic cleaning solution may become entrapped beneath the protective sheath and potentially infect or injure the patient. Loss of electrical integrity also increases the chance of heat generation. Early scopes had a high incidence of electrical failure. However, with newer scopes this problem has been considerably reduced (although not totally eliminated). Continual monitoring of electrical integrity is recommended. In our own laboratory, electrical safety is checked after each examination (the cleaning solution in which the scope is immersed is used as the test bath). Bacteriologic safety. The scope should be cleaned after each use and immersed in a gluteraldehyde solution for 10 to 20 minutes. 1 More prolonged immersion may actually cause increased deterioration of the scope. Adherent gluteraldehyde solution should be washed off and allowed to dry before reuse. Careful visual inspection and electrical checks

are necessary to detect rents in the scope skin that may retain caustic cleaning solution or potentially infective organisms. Bacteremia and risk of endocarditis with upper endoscopy54-58 or TEE 59-61 are considered negligible. Multiple studies suggest that routine prophylaxis is unnecessary. However, a minority of investigators have suggested prophylaxis. 62 •63 In a small subset of high-risk patients, such as those with prosthetic valves, prior endocarditis, or conduits, administering prophylaxis may be clinically justified.

DISCUSSION

Complications are avoided by proper training in indications and contraindications and introduction of the transesophageal scope. 1 Initial training should include supervised introduction of the scope under the tutelage of a gastroenterologist or experienced cardiologist who performs large numbers of transesophageal examinations. Prerequisite echocardiographic training should include a minimum level II (6 months training and subsequent close supervision of interpretation of results). To master new tomographic anatomy, interpretation of unique transesophageal images and a minimum of 30 to 50 personally performed transesophageal examinations are considered minimal training. Training of anesthesiologists and other subspecialists would place equal emphasis on knowing the basics of echocardiography as well as TEE. Once this technique is mastered, between 75 to 100 examinations per year are needed to maintain competency. Individuals or laboratories

Volume 5 Number 3 May-June 1992

performing small numbers of transesophageal examinations should do them only under direct supervision and in close collaboration with an established endoscopy laboratory. Most reported complications have occurred during the actual transesophageal examination as opposed to the introduction of the probe. Because the examination is directed toward cardiac patients, appropriate monitoring and availability of resuscitation personnel and equipment is considered accepted practice. Electrocardiogram, pulse oximetry, blood pressure, oxygen, suction apparatus, and observation of the patient by trained personnel are essential for continuous monitoring of cardiovascular status. The potential complications of any drug used during the transesophageal procedure must be understood and its administration monitored. The limitations and pitfalls of TEE can be best minimized by experience. Initial training should not be circumvented, and maintenance of competency should be strictly monitored. Physicians with less than level II echocardiographic training should work in close collaboration with an active echocardiographic laboratory and receive appropriate review of ongoing examinations. Individual as well as laboratory standards for maintenance of competency should be established. Transesophageal scopes are being continuously modified and refined to address current limitations. Scopes allowing multiple-plane imaging and continuous-wave Doppler and with small tip and shaft size are now being introduced. Display and safety features continue to improve. An ideal scope should incorporate all of the functions of currently available surface transducers and have a wider field of view and smaller diameter.

CONCLUSION

As with any technology, there are limitations and pitfalls. Invasive TEE also has its potential for serious complications. Because of the new presentation of cardiac and extracardiac anatomy, unfamiliar but normal anatomy may initially be confused with abnormal. Additionally, certain structures are viewed in a manner that mimic pathologic conditions. Because of the superior resolution afforded by TEE, phenomena such as spontaneous contrast and ghosting are much more commonly observed when compared with transthoracic imaging. Highly detailed anatomy, such as atrial muscle bundles, sutures, and epicardial fat, are structures to be recognized and differentiated from thrombus, vegetations, or mass. Al-

Critical appraisal of TEE 303

though TEE has been a dramatic step forward in diagnostic imaging, there is potential for serious misinterpretation. This article discusses most of these potential problems; however, there will always be unique situations that must be consistently addressed and differentiated as to normal, artifact, new observation, or misinterpretation.

REFERENCES l. Seward JB, Khandheria BK, Abel MD, eta!. Transesophageal echocardiography: technique, anatomic correlations, implementation, and clinical applications. Mayo Clin Proc 1988;63:649-80. 2. Seward JB, Khandheria BK, Edwards WD, et a!. Biplanar transesophageal echocardiography: anatomic correlations, image orientation, and clinical applications. Mayo Clin Proc 1990;65:1193-213. 3. Daniel WG, Erbel R, Kasper W, et a!. Safety of transesophageal echocardiography. A multicenter survey of 10,419 examinations. Circulation 1991;83:817-21. 4. Khandheria BK, Seward JB, Bailey K, et a!. Safety of transesophageal echocardiography: experience with 2070 consecutive procedures [Abstract]. JAm Coli Cardiol1991; 17:20A. 5. Geibel A, Kasper W, Behroz A, Przewolka U, Meinertz T, Just H. Risk of transesophageal echocardiography in awake patients with cardiac disease. Am J Cardiol 1988;62:337-9. 6. Savino JS, Weiss S. Safety of transesophageal echocardiography is still unclear (Letter). Anesthesiology 1990;73: 366-7. 7. Hanrath P, Bleifeld W, Souquet J, eds. Cardiovascular diagnosis by ultrasound: transesophageal, computerized, contrast, Doppler echocardiography. The Hague: Martinus Nijhoff, 1982. 8. Erbel R, Khandheria BK, Brennecke R, Meyer J, Seward JB, Tajik AJ, eds. Transesophageal echocardiography. Berlin: Springer-Verlag, 1989. 9. Cyran SE, Kinball TR, Meyer RA, et a!. Efficacy of intraoperative transesophageal echocardiography in children with congenital heart disease. Am J Cardiol1989;63:594-8. 10. Cyran SE, Myers JL, Gleason MM, Seber HS, Baylen BG, Waldhausen JA. Application of intraoperative transesophageal echocardiography in infants and small children. J Cardiovasc Surg (Torino) 1991;32:318-21. 11. Ritter SB. Transesophageal real-time echocardiography in infants and children with congenital heart disease. J Am Coli Cardiol1991;18:569-80. 12. Ofili EO, Rich MW. Safety and usefulness of transesophageal echocardiography in persons aged greater than or equal to 70 years. Am J Cardiol1990;66:1279-80. 13. Flachskampf FA, Hoffinann R, Hanrath P. Experience with a transesophageal echotransducer allowing full rotation of the viewing plane: the omniplane probe [Abstract]. JAm Coli Cardiol1991;17:34A. 14. Omoto R, Kyo S, Matsumura M, eta!. Recent technological progress in transesophageal color Doppler flow imaging with special reference to newly developed biplane and pediatric probes. In: Erbel R, et a!., eds. Transesophageal echocardiography. Berlin: Springer-Verlag, 1989:21-6. 15. Seward JB, Khandheria BK, Tajik AJ. Wide-field transesophageal echocardiographic tomography: feasibility study. Mayo Clin Proc 1990;65:31-7.

304

Seward et al.

16. Tsui-Lieh Hsu, Weintraub AR, Ritter SB, Pandian NG. Panoramic transesophageal echocardiography. Clinical application of real-time, wide-angle, transesophageal two-dimensional echocardiography and color-flow imaging. Echocardiography 1991;8:677-85. 17. Skotton DJ, Melton HE Jr, Pandian NG, eta!. Detection of acute myocardial infarction in closed chest dogs by analysis of regional 2-dimensional echocardiographic gray-level distributions. Circ Res 1983;52:36-44. 18. Chandrasekaran K, ChuA, GreenleafJF, eta!. 2D echo quantitative texture analysis of acutely ischemic myocardium. Circulation 1986;74:II-27l. 19. Fitzgerald PJ, McDaniel MD, Rolett EL, Strohbehn JW, James DH. Two-dimensional tissue characterization: backscatter power, endocardial wall motion and their phase relationship for normal, ischemic and infarcted myocardium. Circulation 1987;76:850-9. 20. Kuroda T, Kinter TM, Seward JB, Yanagi H, GreenleafJF. Accuracy of three-dimensional volume measurement using biplane transesophageal echocardiography probe: in vitro experiment. J AM Soc ECHOCARDIOGR 1991;4:475-84. 21. Wollschlager H, Zeiher AM, Klein H-P, Kasper W, Wollschlager S, Just H. Transesophageal echo computer tomography: a new method for dynamic 3-D imaging of the heatt [Abstract]. JAm Coli Cardiol1989;13(suppl A):68A. 22. Omoto R, Kyo S, Matsumura M, et a!. New direction of biplane transesophageal echocardiography with special emphasis on real-time biplane imaging and matrix phased-array biplane transducer. Echocardiography 1990;7:699. 23. Urbanowicz JH, Kemoff RS, Oppenheim G, Parnagian E, Billingham ME, Popp RL. Transesophageal echocardiography and its potential for esophageal damage. Anesthesiology 1990;72:40-3. 24. O'Shea JP, Southern JF, D'Ambra MN, eta!. Effects of prolonged transesophageal echocardiographic imaging and probe manipulation on the esophagus-an echocardiographicpathologic study. JAm Coli Cardiol1991;17:1426-9. 25. Cucchiara R, Nugent M, Seward J, Messick J. Air embolism in upright neurosurgical patients. Detection and localization by two-dimensional transesophageal echocardiography. Anesthesiology 1984;60:353-5. 26. Dewhirst WE, Stragand JJ, Fleming BM. Mallory-Weiss tear complicating intraoperative transesophageal echocardiography in a patient undergoing aortic valve replacement. Anesthesiology 1990;73:777-8. 27. Watts HD. Mallory-Weiss syndrome occurring as a complication of endoscopy. Gastrointest Enclose 1976;22:171-2. 28. Knauer CM. Mallory-Weiss syndrome. A characterization of 75 Mallory-Weiss lacerations in 528 patients with upper gastrointestinal hemorrhage. Gastroenterology 1976;71:5-8. 29. Seward JB, Tajik AJ. Transesophageal echocardiography in congenital heatt disease. Am J Cardiac Imaging 1990;493:215-22. 30. Fyke FE, Tajik AJ, Edwards WD, Seward JB. Diagnosis of lipomatous hypertrophy of the atrial septum by two-dimensional echocardiography. JAm Coli Cardiol1983;1:1352-7. 31. Hanley PC, Tajik AJ, Hynes JK, eta!. Diagnosis and classification of atrial septal aneurysm by two-dimensional echocardiography: report of 80 consecutive cases. J Am Coli Cardiol1985;6:1370-82. 32. Schluter M, Langenstein BA, Thier W,et a!. Transesophageal two-dimensional echocardiography in the diagnosis of cor triatriatum in the adult. JAm Coli Cardiol1983;2:1011-5. 33. Goldfarb A, Weinreb J, Daniel WG, Krown I. A patient with

Journal of the American Society of Echocardiography

right and left atrial membranes: the role of transesophageal echocardiogaphy and magnetic resonance imaging in diagnosis. JAM Soc ECHOCARDIOGR 1989;2:350-3. 34. Freedberg RS, Weinreb J, Gluck M, Kronzon I. Paraesophageal hernia may prevent cardiac imaging by transesophageal echocardiography. J AM Soc EcHOCARDIOGR 1989;2:202-3. 35. Yoshida K, Yoshikawa J, Yamaura Y, eta!. Assessment of mitral regurgitation by biplane transesophageal color Doppler flow imaging. Circulation 1990;2:1121. 36. Castello R, Pearson AC, Lenzen P, Labovitz AJ. Effect of mitral regurgitation on pulmonary venous velocities derived from transesophageal echocardiography color-guided pulsed Doppler imaging. JAm Coli Cardiol1991;17:1499-506. 37. Khandheria BK, Seward JB, Oh JK, eta!. Valve and limitatons of transesophageal echocardiography in assessment of mitral valve prosthesis. Circulation 1991;83:956-68. 38. Daniel WG, Nellessen U, Schroder E, eta!. Left atrial spontaneous echo contrast in mitral valve disease: an indicator for an increased thromboembolic risk. J Am Coli Cardiol 1988;11:1204-11. 39. Black IW, Hopkins AP, Lee LL, Walsh WF, Jackobson BM. Left atrial spontaneous contrast: a clinical and echocardiographic analysis. JAm Coli Cardiol 1991;18:398-404. 40. Erbel R, Rennollet H, Engberding R, Visser C, Daniel W, Roelandt J, (European Cooperative Study Group for Echocardiography). Complementary role of echocardiography in the diagnosis of aortic dissection including transesophageal echocardiography. Lancet 1989;1:457/61. 41. Karalis DG, Chandrasekaran K, Victor MF, Ross JJ Jr, Mintz GS. Recognition and embolic potential of intraaortic atherosclerotic debris. JAm Coli Cardiol 1991;17:73-8. 42. Kronzon I, Demopoulos L, Schrem SS, Pasternack P, McCauley D, Freedberg RS. Pitfalls in the diagnosis of thoracic aortic aneurysm by transesophageal echocardiography. JAM Soc ECHOCARDIOGR 1990;3:145-8. 43. Shamir M, Schuman B. Complications of fiberoptic endoscopy. Gastrointest Enclose 1980;26:86-91. 44. Silvis SE, Nebel 0, Rogers G, Sugawa C, Mandelstam P. Endoscopic complications. JAMA 1976;9:928-30. 45. Miller G. Komplikationen bei der Endoskopie des oberen Gastrointestinaltraktes. Leber Magen Darm 1987;5:299304. 46. Oh JK, Seward JB, Khandheria BK. Transesophageal echocardiography in critically ill patients. Am J Cardiol 1990;66: 1492-5. 47. Pearson AC, Castello R, Labovitz A]. Safety and utility of transesophageal echocardiography in the critically ill patient. Am Heart J 1990;119:1083-9. 48. Fagan LF, Weiss R, Castello R, Labovitz AJ. Transtracheal placement and imaging with a transesophageal echocardiographic probe. Am J Cardiol1991;67:909-10. 49. Silvey SV, Staughton TL, Pearl W, Collaw WA., Belbel RJ. Rupture of the outer partition of aortic dissection during transesophageal echocardiography. Am J Cardiol 1991; 68:286-7. 50. Messina AG, Paranicas M, Fiamengo S, et a!. Risk of dysphagia after transesophageal echocardiography. Am J Cardiol 1991;67:313-4. 51. O'Shea JP, D'Ambra MN, Magro C, eta!. Transesophageal echocardiography: is it safe to the esophagus? An in-vivo study [Abstract]. Circulation 1988;78(suppl II):II-440. 52. Urbanowicz JH, Kernoff RS, Oppenheim G, Parnagian E, Billingham ME, Popp RL. Transesophageal echocardiogra-

Volume 5 Number 3 May-June 1992

53. 54. 55. 56. 57.

58.

phy and its potential for esophageal damage. Anesthesiology 1990;72:40-3. Messina AG, Paranicas M, Fiamengo S, Yao FS, Krieger K, Isom OW, Devereux RB. Risk of dysphagia after transesophageal echocardiography. Am J Cardiol 1991;67:313-4. Botoman VA, Surawicz CM. Bacteremia with gastrointestinal endoscopic procedures. Gastrointest Enclose 1986;32:342-6. Leitch DG, Collins JSA, Radhakrishnan S, O'Neill G, Dannatt J. Bacteremia following endoscopy. Br J Clin Pract 1986;40:341-2. Norfleet RG, Mitchell PD, Mulholland DD, Philo]. Does bacteremia follow upper gastrointestinal endoscopy? Am J Gastroenterol 1981;76:420-2. Perucca PJ, Meyer GW. Who should have endocarditis prophylaxis for upper gastrointestinal procedures. Gastrointest Endosc 1985;31:285-7. Shorvon PJ, Eykyn SJ, Cotton PB. Gastrointestinal instrumentation, bacteremia, and endocarditis. Gut 1983; 24:1078-93.

Critical appraisal of TEE

305

59. Steckelberg JM, Khandheria BK, Anhalt JP, eta!. Prospective evaluation of the risk of bacteremia associated with transesophageal echocardiography. Circulation 1991;84:177-80. 60. Melendez LJ, Chan KL, Cheung PK, Sochowski RA, Wong S, Austin TW. Incidence of bacteremia in transesophageal echocardiography: a prospective study of 140 consecutive patients. JAm Coli Cardiol1991;18:1650-4. 61. Voller H, Schroder KM, Gast D, et a!. Does the incidence of positive blood cultures during transesophageal echocardiography necessitate antibiotic prophylaxis [Abstract]? Circulation 1990;82(suppl III):III-244. 62. Gorge G, Erbel R, Henrichs KJ, Wenchel H, Werner H, Mayer J. Positive blood cultures during transesophageal echocardiography. Am J Cardiol1990;65:1404-5. 63. Foster E, Kusumoto FM, Sobol SM, Schiller NB. Streptococcal endocarditis temporally related to transesophageal echocardiography. JAM Soc ECHOCARDIOGR 1990;3:424-7.