Article Text

Transesophageal Dobutamine Stress Echocardiography With Tissue Doppler Imaging for Detection and Assessment of Coronary Artery Disease
  1. Antonio Vitarelli,
  2. Ysabel Conde,
  3. Marco Ferro Luzzi,
  4. Giulia Di Benedetto,
  5. Rossella Giubilei,
  6. Tiziana Leone,
  7. Ester Cimino
  1. From the Department of Cardiology, “La Sapienza” University, Rome, Italy.
  1. Address correspondence to: Antonio Vitarelli, MD, Via Lima 35, 00198 Rome, Italy. E-mail: vitar{at}


Background Transesophageal dobutamine stress echocardiography (T-DSE) has been shown to be a sensitive and specific technique for the detection of myocardial ischemia. A major limitation of echocardiographic study interpretation, however, is the subjective visual analysis of endocardial motion and wall thickening, which is only semiquantitative.

Methods To analyze whether T-DSE with the use of tissue Doppler imaging (TDI) during graded dobutamine infusion may be useful to detect and quantify stress-induced myocardial ischemia by changes in myocardial velocities, 70 patients undergoing coronary arteriography were studied with T-DSE and TDI. Midesophageal and transgastric short- and long-axis images were obtained at each level of dobutamine infusion. T-DSE was successful in 67 patients (96%). Baseline resting pulsed and color peak systolic (S) and early diastolic (E) velocities of the anterior, septal, lateral, and inferior walls were examined.

Results Pulsed and color TDI correlated well at rest and after stress. Fifteen patients had a normal response to dobutamine, and 52 patients had inducible ischemia by two-dimensional criteria. In the normal group, there was a significant dose-dependent increase in S and E velocities. Compared with those in the normal group, patients with coronary artery disease (CAD) had lower resting S and E velocities and blunted S wave increase or E wave decrease during DSE.

Conclusions T-DSE with TDI is a feasible and accurate test for the quantitative assessment of patients with CAD who have impaired augmentation of systolic and diastolic myocardial velocities during dobutamine infusion.

  • transesophageal echocardiography
  • dobutamine stress test
  • coronary artery disease
  • imaging
  • diagnosis

Statistics from



Dobutamine stress echocardiography has been shown to be a sensitive and specific technique for the detection of myocardial ischemia.1-5It may also provide accurate information in risk stratification after myocardial infarction,6,7preoperative risk assessment,8and identification of “stunned” and “hibernating” myocardium.9,10

However, the technique is limited in that transthoracic echocardiographic imaging is poor in some patients and not of sufficient quality to allow accurate delineation of endocardial boundaries. This is most commonly seen in obese patients or in patients with obstructive pulmonary disease or chest wall deformities. Transesophageal echocardiography with improved visualization of epicardial and endocardial borders can overcome some of the limitations inherent in transthoracic echocardiography10-14and has been used to monitor patients at high risk for intraoperative ischemia.15,16

A major limitation of echocardiographic study interpretation is the subjective visual analysis of endocardial motion and wall thickening, which is only semiquantitative. Tissue Doppler imaging (TDI) is a novel echocardiographic technique that can be used to quantitatively assess low-velocity motion of myocardial walls with excellent temporal resolution.17-20The aim of this study was to analyze whether multiplane transesophageal echocardiography with the use of TDI during graded dobutamine infusion may be useful to detect and quantify stress-induced myocardial ischemia by changes in myocardial velocities.


Study Patients

Seventy subjects (61 male and nine female) were recruited from patients undergoing elective cardiac catheterization for evaluation of chest pain. Exclusion criteria included 1) history of recent myocardial infarction; 2) unstable angina; 3) history of ventricular tachycardia or ventricular fibrillation; 4) current atrial fibrillation, atrial flutter, or multifocal atrial tachycardia; 5) clinical evidence of significant valvular heart disease by echocardiography; and 6) history or evidence of mediastinal, pulmonary, or esophageal mass. All patients were in sinus rhythm, and 22 had a history of myocardial infarction (anterior in 11, posterolateral in two, and inferior in nine). In all patients, antianginal medications were continued during the day of the study.


Transesophageal two-dimensional (2D) echocardiograms were recorded with commercially available instruments (Toshiba Power Vision 8000, Toshiba Corporation, Tokyo, Japan). Both standard echo and TDI techniques were used, including wall motion score and regional myocardial velocities, in the longitudinal and circumferential direction.

TDI is a modification of routine color flow Doppler signal processing, bypassing the high-pass filter and inputting the comparatively lower frequency Doppler data from myocardial motion directly into the autocorrelator. Calculated velocity data are color-coded and superimposed on the conventional 2D images (Figure 1A). Tissue Doppler images were recorded using a multicolored postprocessing velocity map that displayed increasing velocity values toward the transducer in shades of red to orange to yellow, respectively, and increasing velocities away from the transducer in shades of blue to turquoise to green, respectively.

Figure 1.

A. Velocity data are color-coded and superimposed on the conventional 2D transesophageal images. B. After positioning the region of interest in the middle of each segment, a tissue velocity profile throughout the cardiac cycle is displayed. C. After color TDI images have been obtained, the same ultrasound device is used to acquire pulsed wave TDI.

The left ventricle could be imaged from the midesophageal and transgastric views in all patients in this study. Tissue velocity values were well below the Nyquist limits of the selected velocity ranges, and aliasing did not occur. A caliper was used to position the region of interest in the middle of each segment, and a tissue velocity profile throughout the cardiac cycle was displayed (Figure 1B).

After color TDI images had been obtained, the same ultrasound device was used to acquire pulsed wave TDI (Figure 1C). Sample volumes were set on the myocardium corresponding to anteroseptal, posterior, posteroseptal, lateral, anterior, and inferior wall in midesophageal and transgastric views. The acoustic power and filter frequencies were set to the lowest values possible, and the sample volumes were set at myocardial wall (width of approximately 8 mm). Baseline resting peak systolic (S) and early diastolic (E) velocities of the anterior, septal, lateral, and inferior walls were recorded at baseline and during stress at a speed of 10 cm/s with simultaneous electrocardiogram (ECG) tracings. Pulsed TDI measurements of peak velocity were made off-line from videotape.


Transesophageal imaging was performed at baseline and during dobutamine infusion. All patients were studied in the left lateral decubitus position. Before passing the transesophageal probe, 1 to 3 mg of intravenous benzodiazepine was administered for sedation and 10% lidocaine spray was used for topical anesthetic. Heart rate and blood pressure were continuously monitored during the procedure. A 12-lead ECG was obtained during each stage of the protocol. All echocardiographic studies were performed using a 5-MHz multiplane transesophageal probe.

At baseline, separate sets of 2D and TDI images were obtained in the midesophageal and transgastric views. The tissue Doppler velocity range was selected at the time of baseline data acquisition to display S velocity in the low- to midportion of the color display range to allow for detection of increases in systolic velocity at peak dobutamine stress within the same velocity range. Proper selection of the velocity range was important because velocity values that exceeded the limits of the selected range were displayed as saturated at the highest color-coded value and could therefore be underestimated.

During dobutamine infusion, the transesophageal probe was left in the transgastric position to monitor for the development of regional wall motion abnormalities. After baseline images were obtained, dobutamine infusion was begun at 10 μg/kg body weight per minute. The infusion was increased by 10 μg/kg per minute at 3-minute intervals up to a maximal infusion rate of 50 μg/kg per minute. If 85% of the predicted maximal heart rate was not attained at 50 μg/kg per minute, then 0.5 mg of atropine was given intravenously.

Transgastric short- and long-axis 2D and TDI images at each level of dobutamine infusion were obtained 2 minutes into each stage. A 12-lead ECG was also obtained 2 minutes into each stage. Criteria for termination of infusion were attainment of 85% of the predicted maximal heart rate, development of new or worsened wall motion abnormality, angina, arrhythmia, ECG changes consistent with ischemia, or intolerance of the transesophageal probe.

At peak dose, images were obtained in the transgastric view, and the probe was then withdrawn into the esophagus for the lower esophageal views. Baseline and peak images were digitized in a dual screen continuous loop format for later review and analysis.

Analysis of Echocardiograms

All echocardiographic studies were interpreted in a blinded manner using a standard 16-segment model.21The development of a new or worsened regional wall motion abnormality (at least two hypokinetic or akinetic segments at peak stress) was considered to be a positive 2D echocardiographic result.

Peak TDI velocities were determined as the means of the measurements obtained during five consecutive beats in systole (S velocity) and early diastole (E velocity). According to the results of previous studies,22,23a ≤2 cm/s stress-induced S increase or a >2 cm/s stress-induced E decrease were used as indicators for ischemic response during dobutamine stress and, thus, for a positive TDI test.

To assess intraobserver variability, one investigator reanalysed the TDI Doppler studies in 10 patients. Ten randomly selected and previously recorded studies were assessed by a second investigator to obtain interobserver reproducibility of the same set of measurements. Results of intra- and interobserver variability were expressed24as the mean difference between observations divided by their average measurement.

Coronary Angiography

All patients underwent cardiac catheterization within 7 days of the study. Biplane angiography was performed using the Sones or the Judkins technique. Two or more projections were obtained for each coronary artery. Significant coronary artery disease (CAD) was defined as >50% angiographic reduction in the luminal diameter of any of the three coronary arteries or their primary branches. The studies were interpreted by experienced angiographers without knowledge of the echocardiographic results.

Statistical Analysis

Data were expressed as mean±SD. The magnitude of the group differences was assessed by Student's t test, and correlations between paired values were obtained by linear regression analysis. Color TDI and pulsed TDI were compared by linear regression. Significance was determined as P<0.05. Agreement between TDI measurements was expressed by Bland-Altman plots.25Sensitivity, specificity, predictive value, and overall accuracy were calculated by using standard formulas.



Transesophageal dobutamine stress echocardiography (T-DSE) was successful in 67 patients (96%). In one patient, the test was terminated before dobutamine infusion because the baseline images were inadequate for interpretation (this patient had a large ventral hernia). Two other patients were excluded from the final analysis because they did not attain the target heart rate. In five patients, 0.5 mg of atropine was given to increase the chronotropic response to each maximal heart rate. Of these patients, one did not attain the target heart rate and was excluded from the final analysis. One other patient did not attain the target heart rate but was not given atropine at the discretion of the operator and was also excluded. Mean study duration from the time of transesophageal intubation to completion of peak stress imaging was 15.6±0.7 minutes.


No patient experienced a complication or adverse outcome because of esophageal intubation. All ECGs immediately after intubation and before dobutamine infusion had no evidence of ischemia. One patient had a brief period of asymptomatic supraventricular tachycardia at peak dobutamine infusion that responded to cessation of the infusion and a small dose of intravenous beta-adrenergic blocking agent.

Coronary Arteriography

One-vessel disease was present in 19 patients, and multivessel disease was present in 36. There were significant LAD lesions in 33 patients, CX in 16 patients, and RCA in 23 patients. A diameter reduction of >75% was observed in 24 patients and of >90% in 18 patients. When normality versus significant stenosis (>50%) was assessed, intraobserver and interobserver agreement both occurred in 96% of the arterial segments evaluated.

Significant CAD was excluded in 12 patients. These patients with normal LV function and normal stress echocardiograms were selected as a control group (10 men, two women, aged 51±9 years).

Conventional Transesophageal Stress Echocardiography

There were a total of 472 segments interpreted to be hypokinetic or akinetic at peak stress by routine 2D visual assessment, 192 normal segments in patients without CAD and 408 normal segments in otherwise abnormal patients.

TDI Availability and Variability

Despite the suboptimal incidence angle of the ultrasound beam for Doppler calculations of lateral segmental velocity, TDI data were available from a high proportion of all segments (Table 1), with the exception of the apical segments from the apical views. The interobserver variability of TDI velocity analysis was 2.9±17.1%, and intraobserver variability was 4.1±8.9%.

Table 1.

TDI data availability (67 patients).

Comparison Between Color TDI and PW TDI

In 67 patients, color and PW TDI were assessed at rest and during the stress test in the six basal segments of each wall (anteroseptal, posterior, lateral, septal, anterior, and inferior) from transesophageal views. Both color and PW TDI were obtainable in 897 segments at rest and 681 at peak stress. PW and color TDI correlated well at rest (S wave: r=0.83, P<0.0001; E wave: r=0.85, P<0.0001) and at peak stress (S wave: r=0.87, P<0.0001; E wave: r=0.88, P<0.0001). As shown in Figure 2, pulsed Doppler systolic values were slightly greater than color TDI values, this difference being 0.91±1.07 cm/s at rest (P<0.001) and 0.22±1.98 cm/s at peak stress (P<0.05). However, because of their optimal correlation, only color TDI values were used for further analysis.

Figure 2.

Bland-Altman plots of differences (y-axis) between color and pulsed waves systolic velocities and the mean of color PW velocity (x-axis) at rest (A) and after dobutamine stress (B).

Response to Dobutamine

Dobutamine increased heart rate in a gradual and consistent fashion in patients with no CAD (Table 2) from 72±15 beats/min at baseline to a peak of 131±16 beats/min (P<0.01). In patients with CAD, heart rate rose from 74±11 beats/min to 116±12 beats/min (P<0.01). In 21 patients, peak heart rate occurred at doses lower than the maximal dose of dobutamine achieved. Systolic blood pressure rose from 136±16 mm Hg to 176±27 mm Hg (P<0.01) in patients with no CAD and from 132±19 to 168±26 (P<0.01) in patients with CAD.

Table 2.

Peak physiologic responses to T‐DSE.

TDI velocity data at maximal stress are summarized in Table 3 and Table 4. Myocardial velocities in 521 normal segments increased by an average of 103% at peak stress. Scarred (akinetic) segments had a lower TDI than normal segments (P<0.001) at rest and stress. Ischemic (hypokinetic) segments were not different from normal segments at rest but had a lower peak stress systolic TDI (Figure 3) than normal segments (5.9±2.8 cm/s vs 9.8±3.1 cm/s, P<0.001). TDI in segments with ischemia or scar were not significantly different at rest (4.7±1.1 cm/s vs 3.8±1.2 cm/s, P=not significant) or at stress (6.1±2.4 cm/s vs 5.5±2.6 cm/s, P=not significant). The normal TDI increment with stress was 4.4±2.8 cm/s in normal segments. This increment was significantly greater than the TDI increment for ischemic segments (1.6±1.9 cm/s, P<0.001) and scar segments (1.5±1.7 cm/s, P<0.001).

Table 3.

Segmental peak systolic and diastolic velocity data at maximal stress (midesophageal views).

Table 4.

Segmental peak systolic and diastolic velocity data at maximal stress (transgastric views).

Figure 3.

Myocardial velocity profile displays through cardiac cycle in normal (A,B) and ischemic (C,D) segment at rest (left) and after dobutamine infusion (right). A blunted response in the ischemic segment is observed compared with the normal velocity increase in the normal segment.

Peak E velocity during diastolic filling showed a stress-induced reduction by >2 cm/s in patients with significant coronary stenosis. In ischemic segments, stress-induced mean decrease was 1.3±2.4 cm/s, and these changes were significant compared with control (P<0.001) and scar segments (P<0.01).

Sensitivity and Specificity

Sensitivity, specificity, and accuracy for 2D echocardiography and TDI S and E velocities (excluding apical segments) to identify CAD at maximal stress appear in Figure 4. The cut-off point in the level of E velocity reduction at 2 cm/s showed a sensitivity of 91%, specificity of 95%, negative predictive value of 78%, and positive predictive value of 97%. The systolic parameter change in S velocity with a cut-off level of ≤2 cm/s had a sensitivity of 84% and a specificity of 77%.

Figure 4.

Sensitivity (SENS), specificity (SPEC), positive predictive value (PPV), negative predictive value (NPV), and accuracy (AC) for 2D, systolic TDI velocities (S-TDI) and diastolic TDI velocities (E-TDI) to identify CAD at maximal stress.


The present study shows that T-DSE with TDI is a feasible and accurate method for the evaluation of patients with CAD with high sensitivity and specificity.

Safety and Feasibility

Transesophageal echocardiography is a semi-invasive technique that requires greater skills than the transthoracic examination and imposes a small risk for serious complications. In our study group, no significant adverse reactions were observed. The side effects relate to dobutamine infusion were self-limited and did not affect the diagnostic accuracy of the test, which is in agreement with previous report data in patients with CAD.3,11-14The procedure was generally well tolerated by patients, with only 1% of the studies terminated because of patient discomfort and intolerance of the probe.

Previous Studies and Present Investigation

The results of the present investigation are in agreement with those of previous studies on stress echocardiography. Overall sensitivity (up to 91%) and specificity (up to 95%) seem to be slightly superior to that of previous transthoracic and transesophageal studies,1,2,11-14although no direct comparison in prospective studies has been accomplished between transthoracic and transesophageal stress images.

The proportion of patients with CAD who need a transesophageal examination for reliable assessment of echocardiographic response to stress varies depending on the operators' skills, the interpreters' experience and the use of videotape or digitizing systems for image analysis. In the patients at our institution, up to 20% have inadequate transthoracic echocardiographic endocardial definition even after the introduction of tissue harmonic imaging.

To avoid the potential pitfalls of subjective interpretation, several techniques have been proposed to quantify myocardial wall motion more objectively. The centerline method requires time-consuming postprocessing.26Quantitative approaches based on backscatter, such as color kinesis and automatic boundary detection,27-31require good image quality in the region of the endocardial border. Recent studies22,23have evaluated the feasibility of applying TDI to dobutamine stress transthoracic echocardiography and determined the normal and abnormal quantitative regional endocardial velocity responses to dobutamine stress.

The use of TDI technique in conjunction with T-DSE is a further application of this use and would be obviously limited to the transesophageal approach. TDI is both quantitative and has better signal-to-noise ratio than 2D imaging and can be displayed as spectral patterns or can be color-coded. PW TDI has the advantage of good spatial resolution at the site of the sample volume and high temporal resolution but requires separate assessment of each myocardial segment. In contrast, color TDI allows simultaneous visualization and data acquisition of multiple myocardial segments in the same echocardiographic view. This technique requires postprocessing to obtain quantitative data. It is also possible to use high frame rates (>70 images/s) and off-line software for simple and rapid postprocessing of color data. Nonetheless, the velocity measurements are not totally equivalent: PW TDI measures true peak velocity, and because color TDI data are autocorrelated, the measured values are closer to modal velocity. Despite these differences, PW and color TDI correlated well at rest and during stress; the differences between measurements were greater at stress than at rest because of reduction in the heart rate between poststress acquisition of color and PW TDI data.

TDI Response to Dobutamine

It has been shown in animal studies19that TDI accurately reflects changes in regional and global left ventricular function induced by dobutamine or esmolol. Ischemia during percutaneous transluminal coronary angioplasty balloon inflation is associated with a marked decrease in TDI and a recovery after balloon deflation.20With dobutamine stress, the velocity of normal segments increases at stress, and ischemic segments have a lower velocity response. However, the workload and the hemodynamic responses to dobutamine stress are much less heterogenous than with exercise stress. TDI in ischemic segments is not different from normal segments at rest, but the velocity increment with stress is significantly blunted compared with that in normal subjects. This result may correspond to the susceptibility of longitudinal subendocardial fibers to ischemia.

In isolated muscle preparations, the velocity of shortening is dependent on preload, afterload, inotropic state, metabolic state (ischemia), and frequency of stimulation. The intact heart is more complex than an isolated muscle preparation. Besides geometric factors (left ventricular volumes), regional variation seems to be an independent predictor of stress TDI.32-35The decreasing velocity gradient from base to apex is a result of the apical systolic movement of the mitral annulus. Regional variations in ventricular myoarchitecture have an independent contribution to stress TDI not reflected in isolated muscle preparations.

It is noteworthy that single-vessel CAD has been documented as difficult to diagnose with conventional interpretation, even if a standard dobutamine stress test is performed. By allowing a quantitative and objective assessment of wall motion, TDI seems to be more sensitive and specific in detecting subtle wall motion changes than a point score method. In addition, the dose of dobutamine or atropine, or both, needed to detect ischemic myocardium may be reduced by using TDI to detect subtle changes.

Diastolic left ventricular abnormalities are sensitive early signs of myocardial ischemia and have the additional advantage of persisting longer than systolic disturbances. Quantification of diastolic myocardial function by TDI during dobutamine transthoracic stress tests has been shown23to be a feasible and accurate technique as well as a sensitive alternative to echocardiographic and scintigraphic imaging techniques for stress tests. The present study shows that the transesophageal approach is also suitable to define the quantitative myocardial response to dobutamine stress in normal and ischemic left ventricular segments and increase the sensitivity of the stress test.


The accuracy of Doppler velocity measurements is dependent on the angle between the Doppler beam and the interrogated surface. In this study, the direction of movement interrogated in the apical views was from base toward apex, the latter being relatively fixed throughout the cardiac cycle. Previous studies have examined this longitudinal motion with PW TDI from the apical view, and the direction of movement in the basal and mid zones corresponds closely to that of the Doppler beam.

TDI measures myocardial velocity relative to the ultrasound beam. This may include myocardial velocity caused by thickening, translational motion of the heart, and motion caused by breathing. The extent to which the velocity reflects translation is unresolved, although the correlation of the TDI technique with assessment of wall motion scoring suggests that the measurement is a useful marker of regional function. Recently, the use of tissue velocity endocardial-epicardial gradient rather than absolute velocities has been proposed to detect myocardial ischemic kinesis,35,36and changes in gradient (which has the advantage of being independent of translation) were significantly different in ischemic and normal segments.

Apical segments are common sites of ischemia, but very low apical velocities have been recorded also in normal subjects, presumably reflecting limited systolic apical excursion; it has been suggested from in vitro experience18that at slower than 10% of Nyquist velocity, the velocity measurements become inaccurate and tend to dither around zero. This currently represents an inherent limitation of the technique and is the reason that the modality will likely be used together with, rather than as a replacement for, subjective interpretation. However, the 12-segment model for pulsed TDI (excluding apical segments) used in recent sudies23did not seem to be inferior to the standard 16-segment 2D model.

Another limitation of this study is the necessity of obtaining a transesophageal echocardiogram, and thus, contraindications of transesophageal echocardiography limit patient selection.



Despite these limitations, our study showed that this test is safe, feasible, and sensitive for the detection of CAD and may be used in selected patients. The combined use of conventional 2D scoring system and TDI has potential to enhance the quantitative approach to interpretation of transesophageal dobutamine stress echocardiograms.


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.