INTRODUCTION — Although transthoracic echocardiography remains the cornerstone of diagnostic cardiac ultrasound, transesophageal echocardiography (TEE) is a valuable complementary tool. As compared with transthoracic echocardiography, TEE offers superior visualization of posterior cardiac structures because of close proximity of the esophagus to the posteromedial heart with lack of intervening lung and bone. This proximity permits use of high-frequency imaging transducers that afford superior spatial resolution (image 1).
TEE displays most of the left ventricle (LV) with definition that is equal or superior to that achieved with transthoracic echocardiography. In particular, the full thickness of the myocardium, including the endocardium with its complex endo-architecture, is seen with clarity. However, most TEE views do not fully display the apical portion of the LV, which is truncated or foreshortened. This shortcoming lowers the sensitivity of TEE for detecting a wall motion abnormality limited to the apex, leads to underestimation of chamber volume, and may result in misinterpretation of global LV systolic function. Importantly, the presence of apical thrombus may be missed on TEE alone [1].
The use of TEE for the evaluation of the LV will be reviewed here. The specific roles for TEE in ischemic heart disease, valvular disease, and aortic pathology are discussed in detail separately. However, a brief section on the emerging role of TEE in patients with LV assist devices is included here. (See "Transesophageal echocardiography in the evaluation of aortic valve disease" and "Transesophageal echocardiography in the evaluation of mitral valve disease".)
LEFT VENTRICULAR ANATOMY — Linear measurements of the LV diameter are best performed in the transgastric views at the midventricular level. Accuracy is enhanced using simultaneous multiplane views of the short and long axis (image 2) [1]. These measurements should be performed at end-diastole and end-systole and can be used to calculate fractional shortening.
LV wall thickness can also be measured in these views at end-diastole to detect LV hypertrophy.
SYSTOLIC FUNCTION — While quantitative methods to measure LV systolic function and ejection fraction by TEE have been described, most TEE evaluation of the LV is qualitative. Studies by anesthesiologists in the operating room have documented the validity of this approach for recognizing hypovolemia, the adverse effects of proximal cross-clamping of the aorta, and the influence of various anesthetic regimens on LV systolic function [2-7]. When evaluating LV function, it is important to consider the patient's overall condition. Volume status, sedating medications, and general anesthesia may alter both systolic and diastolic function through their effects on preload and afterload.
An overview of tests to evaluate LV systolic function is presented separately. (See "Tests to evaluate left ventricular systolic function".)
Ejection fraction — Calculation of end-diastolic and end-systolic volumes by TEE using the biplane method of discs (ie, Simpson's rule) has generally been shown to underestimate LV size [8,9]. However, one study was able to demonstrate a close correlation between TEE and transthoracic echocardiographic measurements of LV volumes and ejection fraction [10]. A non-volumetric method has been described that accurately estimates LV ejection fraction using the descent of the mitral annulus toward the apex as an index of the shortening of the ventricular long axis [11].
Estimates of LV function (global and segmental) should also include a transgastric short-axis view. Although this view must be used with some degree of caution because it excludes the longitudinal motion of the heart (ie, basal descent), it remains a reliable approach for judging LV systolic function. With multiplane and three-dimensional TEE, this view can be supplemented with long-axis gastric views that approach the LV apex in some patients, and reliably image the basal inferior wall. In a study comparing several two-dimensional echocardiographic methods with three-dimensional echocardiography [12], the ejection fraction obtained using a single-plane method of discs had the best correlation with three-dimensional measurements. In another intraoperative study of 63 patients, three-dimensional echocardiography was feasible approximately 95 percent but provided no clear advantage over two-dimensional methods [13].
Fractional area change — In lieu of LV ejection fraction, the fractional area change or area ejection fraction (normal >36 percent) has been used, mainly for research purposes, for quantifying changes in the short-axis image of the LV. Using manual planimetry of the area circumscribed by the endocardium at end-diastole (EDA) and end-systole (ESA), the fractional area change can be calculated according to the formula:
Fractional area change = (EDA - ESA) / EDA
The validity of utilizing the fractional area change was evaluated in one study that compared radionuclide blood pool imaging with TEE area ejection fraction in a group of 10 patients in the postsurgical intensive care unit [14]. With LV ejection fractions below 45 percent, the correlation between the reference standard blood pool image and the short-axis measured by planimetry was linear; the correlation was poorer when ejection fractions were normal or above normal. However, discriminating among high ejection fractions is of minor clinical importance, and, as a result, use of the fractional short-axis area change can be considered a valid approach. Fractional area change has also been corrected for afterload using the following formula [15]:
FAC afterload corrected (FACac) = FAC x log10([MAP - RAP] / CI) x 100%
Studies applying this method in comparison with pressure-volume loops intraoperatively have validated this method as being relatively independent of loading conditions.
In addition, if one considers that the descent of the cardiac base represents longitudinal muscle function and is among the first properties to be altered in patients with systolic dysfunction, it is to be expected that an index that is blind to longitudinal changes in cardiac dimensions would be less effective once that vector of motion had been suppressed. The reproducibility of measuring the area and fractional change of the short-axis appears to be acceptable for clinical and investigational use [16,17].
Strain analysis — The two clinically applicable methods for measuring myocardial deformation and deriving indices of systolic function are tissue Doppler and speckle tracking. These methods have been applied to transthoracic as well as transesophageal imaging. Available TEE studies have primarily been performed in the intraoperative setting [18]. In one study, tissue Doppler in 19 intraoperative patients was successful in measuring radial strain and strain rate in the transgastric views in contrast to the speckle-tracking method, which was less successful. However, a second study demonstrated that speckle tracking measurements of radial and longitudinal strain were more successful at differentiating between perfused and non-perfused myocardial segments when compared with visual estimation of wall motion [19]. Normal values for global strain rate have not been established by transesophageal echocardiography. Normal values for global longitudinal strain using various techniques applied to transthoracic echocardiography range between -16 and -19 percent [20,21]. Normal radial strain values are between 45 and 55 percent [22]. (See "Tissue Doppler echocardiography".)
Doppler echocardiography — Doppler evaluation of LV systolic performance is also feasible. Cardiac output has been measured from the pulmonary artery [23,24] and LV outflow tract [25] with the use of velocity-time integrals obtained by pulsed-wave Doppler and continuous-wave Doppler [26] combined with outflow tract diameter estimates of area. Other flow signals, such as transmitral inflow, may also provide cardiac output information. Tissue Doppler measurements of peak systolic myocardial velocity using the basal lateral wall were validated as a measure of systolic LV function against intraoperative pressure-volume loops [27]. Unlike dP/dt derived from the mitral regurgitant jet, the peak systolic myocardial velocity was relatively independent of loading conditions.
Evaluation of regional wall motion — Abnormalities in wall motion associated with myocardial ischemia or infarction are characterized by diminished or absent inward endocardial motion and by impaired systolic myocardial thickening. TEE is sensitive for the detection of acute ischemia or prior infarction and is therefore useful for assessment of regional wall motion as well as global LV function.
As an example, the continuous high-quality imaging of the LV afforded by TEE during surgical procedures makes it ideally suited for the early detection of ischemia [28,29]. Intraoperative changes in wall motion are more predictive of postoperative ischemia or infarction than ECG changes or hemodynamic abnormalities as documented with Swan-Ganz catheter measurements, such as an increase in pulmonary capillary wedge pressure [28,30,31]. In a study of 98 patients studied before and after the induction of anesthesia prior to coronary artery bypass graft surgery, ischemia as defined by TEE was associated with a small elevation in PCWP (3.5 mmHg); however, a rise in PCWP itself had a low sensitivity and predictive value for myocardial ischemia [30].
Views used for evaluation — The transverse and longitudinal plane transgastric views are the most widely used for evaluating wall motion abnormalities. Available three-dimensional echocardiographic matrix array transducers can provide simultaneous visualization of multiple planes, which can facilitate rapid assessment of wall motion. The standard transesophageal LV short axis view (horizontal plane) is obtained at the level of the papillary muscle tips. At this level, the myocardium is supplied by all three major coronary vessels and can be divided into four or six segments for standardization purposes: anterior, anterolateral, inferolateral, inferior, inferoseptal, and anteroseptal (figure 1A-B).
Continuous imaging of this cross-sectional view forms the basis of monitoring wall motion before and immediately after coronary artery bypass surgery and in high-risk patients undergoing non-cardiac surgery. The longitudinal transgastric view, as well as the esophageal views, provide incremental value in detecting additional ischemic segments [32,33].
Measurement of wall motion abnormalities — The most widely employed method for measuring segmental wall motion uses qualitative analysis of endocardial wall motion and segmental thickening (figure 1A-B). By convention, segment systolic wall thickening and endocardial inward motion are usually graded as [34,35]:
●Normal
●Hypokinetic (ie, reduced and delayed contraction)
●Akinetic (ie, absence of inward motion and thickening)
●Dyskinetic (ie, systolic thinning and outward systolic endocardial motion)
Akinesia and dyskinesia usually are the result of a prior myocardial infarction and therefore often reflect nonviable myocardium. Hypokinetic segments, which may result purely from ischemia but may also reflect a prior non-transmural myocardial infarction, generally have at least some viability and often reflect hibernating myocardium [36]. (See 'Limitations' below and "Pathophysiology of stunned or hibernating myocardium".)
Limitations — There are some potential limitations to utilizing TEE for the assessment of regional wall motion:
●In patients with conduction abnormalities, such as left bundle branch block, or ventricular pacing, it can be difficult to distinguish the incoordinate contractions resulting from delayed activation from those of an ischemic wall motion abnormality.
●Regional wall motion abnormalities are not exclusive to ischemia, since they can also be seen with a nonischemic cardiomyopathy as well as myocarditis. In addition, myocardial stunning may cause persistent wall motion abnormality in an adequately revascularized region. It can be difficult to differentiate this finding from ischemia due to inadequate revascularization. (See "Clinical syndromes of stunned or hibernating myocardium".)
●Other possible influences that confound the recognition of acute ischemia by TEE include volume changes and alterations in transducer position and/or angulation, which create de novo unmasking of pre-existing wall motion abnormalities.
●Apical wall motion abnormalities and, potentially, LV thrombus, may be missed by TEE due to failure to adequately visualize the LV apex.
DIASTOLIC FUNCTION — Diastole includes a period of active relaxation during isovolumic relaxation and rapid early filling, a period of passive filling during diastasis, and active filling during atrial contraction. The first phase of early relaxation is energy dependent, the second phase of passive filling is dependent upon ventricular compliance, and the third phase during atrial contraction is highly dependent upon the LV diastolic pressure at the onset of atrial contraction. Impaired LV relaxation, poor LV compliance, or both may be the pathophysiologic mechanisms leading to pulmonary congestion and clinical symptoms of heart failure in patients with diastolic dysfunction [37]. The rates of relaxation and the filling characteristics are dependent upon preload, afterload, heart rate, and contractility [38]. (See "Pathophysiology of heart failure with preserved ejection fraction" and "Echocardiographic evaluation of left ventricular diastolic function in adults".)
Mitral inflow filling pattern — Mitral inflow, alone and in combination with pulmonary venous flow, is the major source of clinically relevant information about diastolic LV behavior. As with transthoracic echocardiography, transmitral inflow is recorded with the Doppler sample volume at the tips of the mitral leaflets from the esophageal view plane that provides the most axial inflow signal (waveform 1A-B) [39,40]. (See "Echocardiographic evaluation of left ventricular diastolic function in adults", section on 'Mitral inflow velocities and isovolumic relaxation time'.)
Important diagnostic parameters derived from the mitral inflow signals include:
●The ratio of peak early filling velocity to atrial filling velocity (E to A ratio)
●The deceleration time of early filling curve
●The isovolumic relaxation time
Two predominant patterns of flow have been recognized on the TEE, reflecting the two major categories of diastolic dysfunction. These are the same as has been seen on transthoracic echocardiography (TTE) (figure 2) [41].
●An impairment in LV relaxation or delayed relaxation is characterized by a decrease in the transmitral E to A ratio (the so-called A-wave dominant pattern), which arises through a combination of prolonged deceleration time and increased contribution of atrial contraction. This form of diastolic dysfunction occurs in LV hypertrophy of any cause and with normal aging. The delayed relaxation inflow pattern usually implies a normal mean filling pressure.
●Diminished or impaired LV filling with elevated diastolic filling pressure is identified by Doppler by "restrictive flow" pattern, with an increased transmitral E to A ratio, a shortened isovolumic relaxation time, and a short deceleration time. This pattern occurs in patients with restrictive cardiomyopathies, such as amyloid heart disease, and in patients with acute and chronic cardiac decompensation associated with a variety of myopathic and pericardial conditions.
As heart failure worsens, a pattern change from delayed relaxation to restriction is often seen. At one point in this progression, as filling pressure rises to pathologic levels, the filling pattern will "pseudonormalize" (figure 2) [41].
In evaluating diastolic function by TEE it is always important to recognize that Doppler transmitral filling patterns are influenced by a variety of factors, including loading conditions, heart rate, pericardial restraint, left atrial pressure and compliance, right and left ventricular interaction, coronary turgor, and intrinsic properties of left atrial and LV muscle as well as abnormal mitral valve function (stenosis and/or regurgitation). As an example, changes in loading conditions during cardiac surgery may induce several changes in filling patterns over the course of a single operation [39].
Pulmonary venous Doppler — Deducing the status of diastolic function through evaluating transmitral inflow patterns is greatly enhanced by also examining the Doppler signals from the pulmonary veins. These signals are particularly accessible and well-resolved when sampled with the aid of TEE/Doppler. Signals obtained in this manner have contributed substantially to our understanding of diastolic hemodynamics. (See "Echocardiographic evaluation of left ventricular diastolic function in adults", section on 'Pulmonary venous flow'.)
TEE demonstrates the entrance of the four pulmonary veins into the left atrium. In the transverse view, the left upper pulmonary vein flow is close to and parallel to the direction of the interrogating beam, and highly resolved color flow Doppler images are readily obtained, helping to guide positioning of the pulsed-wave Doppler sample volume into the proximal 1 cm of the vein. The right upper pulmonary vein is close to and parallel to the direction of the interrogating beam when observed at the vertical (90 degree) view.
Flow in the pulmonary veins is triphasic (image 3A-D) [42-44]:
●The systolic phase is predominant, accounting for more than 55 percent of the total flow integral. The systolic phase consists of two components. The first component results from active atrial relaxation and the second from systolic descent of the cardiac base [45,46].
●The second phase occurs in early diastole and accounts for approximately 40 percent of the total forward inflow. This phase results from ventricular relaxation.
●The third phase occurs in late diastole and is due to atrial contraction. The atrial phase is retrograde and usually small.
Frequently measured variables from pulmonary venous flow velocity tracings with demonstrated utility include:
●Peak systolic and peak early diastolic flow velocities
●Peak velocities of flow reversal at atrial contraction
●Velocity-time integrals of the systolic, early diastolic, and atrial contraction phases
The systolic velocity-time integral (S) is measured from the onset of forward flow following the peak R wave on the electrocardiogram to the point at which it reaches zero flow velocity. The early diastolic velocity-time integral (D) is measured from the onset of the second wave to its crossover with the zero-line. The systolic fraction is equal to S/(S + D). The velocity-time integral of the late diastolic A wave (retrograde flow during atrial contraction) is measured from its onset to the end of negative flow.
In conditions that elevate LV filling patterns, systolic dominance of inflow ceases and its contribution (measured from its VTI) becomes less than 50 percent of total inflow. Concurrently, retrograde atrial flow may increase and its duration exceeds forward atrial transmitral valve [47]. However, in a study of 102 patients, there was poor correlation between the systolic fraction of pulmonary venous flow, the mitral E/A ratio, and the pulmonary capillary wedge pressure [48]. In a regression analysis, the limits of agreement exceeded 10 mmHg, rendering this method unreliable for estimating filling pressures intraoperatively. Additionally, there are numerous other conditions that impact pulmonary venous flow patterns including atrial fibrillation and severe mitral regurgitation, in which systolic flow reversal has been described as one of the signs of severity [49].
Transmitral inflow and pulmonary vein flow patterns are complementary [26]. As an example, restrictive mitral inflow is characterized by a short isovolumic relaxation time, a normal or slightly elevated peak velocity, a short deceleration time, and a low A wave velocity. These features result in an increased E to A ratio and a pattern known as "pseudonormalization" (figure 2) [41]. A restrictive pattern on mitral inflow and decreased systolic fraction on pulmonary vein inflow permit confident recognition of elevated filling pressure. A prolonged retrograde pulmonary venous A wave and a shortened A wave on mitral inflow also confirm elevated pressure.
Tissue Doppler measurements — Mitral e' and a' velocities can be measured during transesophageal echocardiography at the lateral annulus. These measurements have been validated and appear to reflect diastolic function independent of loading conditions. A study performed in the intensive care unit using both TEE and TTE showed that the sensitivities and specificities of estimating PCWP of 15 mmHg or higher were, respectively, 86 and 81 percent for lateral E/E' above 7.5 and 76 and 80 percent for medial E/E' above 9. After volume expansion, the E/E' tracked the changes in wedge pressure [50].
Integrated approach to diastolic assessment — Flow propagation velocity (Vp) can also be examined using color flow Doppler to examine early mitral inflow. Impaired relaxation as defined by this technique is <45 cm/second (image 3A-D) [51].
THE ROLE OF TEE IN PLACEMENT AND MANAGEMENT OF LV ASSIST DEVICES — Circulatory LV assist devices (LVADs) are primarily used in adult patients with end-stage heart failure (stage D) in the setting of severe LV systolic dysfunction. They may be used as a bridge to transplantation, a bridge to recovery of ventricular dysfunction, or increasingly, as destination therapy. There are a number of devices which all provide continuous flow from the LV to the ascending aorta. TEE is usually performed during the operative insertion of the device. The most common LVAD used in the United States is the HeartMate II, which consists of an inflow cannula in the apex of the LV, the pump, and an outflow cannula to the ascending aorta. On the preimplantation examination, it is important to exclude the presence of thrombi in LV apex and left atrium, to evaluate valvular function, and to examine for the presence of an intracardiac shunt. Postimplantation, the position of the inflow cannula within the LV is evaluated to ensure that it is perpendicular to the mitral annulus and not abutting a wall. It is important to insure adequate but not excessive decompression of the LV. The flow rate of the LVAD should be adjusted such that the position of the intraventricular septum is midline and the aortic valve opens only every third or fourth cycle. The degree of aortic regurgitation should be reassessed and, if significant, aortic valve repair or replacement may be required [52,53]. In general, TTE will be adequate in the subsequent follow-up of these patients. Indications for TEE in patients with implanted LVADs include suspected device infection or thrombus.
Use of echocardiography in the emergency care of adults with mechanical circulatory support devices is discussed separately. (See "Emergency care of adults with mechanical circulatory support devices".)
INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5th to 6th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10th to 12th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.
Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)
●Basics topic (see "Patient education: Echocardiogram (The Basics)")
SUMMARY AND RECOMMENDATIONS
●Transesophageal echocardiography (TEE) displays most of the left ventricle (LV) with definition that is equal or superior to that achieved with transthoracic echocardiography. However, most TEE views do not fully display the apical portion of the LV, which is truncated or foreshortened. This shortcoming lowers the sensitivity of TEE for detecting a wall motion abnormality limited to the apex, leads to underrepresentation of chamber volume, and may result in misinterpretation of global LV systolic function. (See 'Introduction' above.)
●While there are well validated quantitative methods to measure LV systolic function and ejection fraction when performing transthoracic echocardiography, most evaluation of the LV by TEE is qualitative. However, qualitative assessment of LV systolic function is generally accurate and has been validated. For the most thorough and accurate estimate of LV systolic function, a transgastric short-axis view of the LV should be obtained in addition to long-axis gastric views that approach the LV apex and demonstrate the basal inferior wall. (See 'Systolic function' above.)
●TEE can be a useful modality for identifying regional wall motion abnormalities of the LV associated with acute ischemia or prior infarction. (See 'Evaluation of regional wall motion' above.)
●Apical wall motion abnormalities and, potentially, LV thrombus, may be missed by TEE due to failure to adequately visualize the LV apex. (See 'Limitations' above.)
●LV diastolic function can also be evaluated with a combination of techniques, including the transmitral inflow pattern, pulmonary vein inflow, and tissue Doppler imaging. (See 'Diastolic function' above.)
●TEE is important during the intraoperative placement of an LV assist device (LVAD) primarily to assess for thrombus, valve dysfunction, and intracardiac shunts preimplantation and for cannula placement postimplantation.
●In patients with indwelling LVADs, TEE may be needed to exclude device infection and thrombus.
