Chapter 22. TEE in the Critical Care Unit

In addition to its role as an intraoperative diagnostic tool, transesophageal echocardiography (TEE) has been shown to be increasingly useful in the critical care setting. Patients in the critical care unit may have impaired cardiac function due to comorbidities or as a result of their critical illness. Indeed, cardiac dysfunction is one of the most common causes of hemodynamic instability and death in critically ill patients.1 Unlike many other diagnostic modalities available to the intensivist, TEE is minimally invasive and can be rapidly performed at the bedside. It can be used to provide valuable information on cardiac function, guide resuscitation and management, and diagnose a wide range of pathologies that may have a negative impact on the critically ill patient (Table 22–1). TEE was initially applied in the intensive care unit (ICU) for the postoperative evaluation of unstable cardiac surgery patients. However, TEE has also been shown to be beneficial in the general ICU setting, changing management in up to two-thirds of cases.2,3 In addition to the fact that the use of echocardiography has a positive impact on the management of patients in the general ICU,4 there is also evidence to support its therapeutic impact and its value in predicting mortality.5 Surprisingly, despite all these advantages, echocardiography is still not yet available in most ICUs, being largely limited by the availability of trained providers.5

The indications for TEE, as previously discussed in this manual, also apply to critically ill patients. The ACC/ AHA/ASE guidelines define hemodynamic instability and suspected aortic dissection as class I indications for echocardiography in the ICU (ACC/AHA/ASE 2003 guidelines) (Table 22–2). Acutely ill trauma patients should also undergo an echocardiographic evaluation if they are suspected to have cardiac tamponade or aortic injury. The most common indication for which TEE is performed in the ICU is hemodynamic instability, followed by endocarditis, assessment of ventricular function, aortic pathology, and miscellaneous others.2

In the general population, transthoracic echocardiography (TTE) is typically the first modality used, as it is noninvasive and more widely available. In the ICU population, however, up to one-half of mechanically ventilated patients cannot be imaged adequately via TTE.6 Barriers to adequate TTE imaging include mechanical ventilation with positive end-expiratory pressure, wounds and dressings, and body habitus. In addition, many ICU and trauma patients may be too unstable to be positioned in the left lateral decubitus position for TTE imaging. In a study comparing TTE to TEE in patients with unexplained hypotension in the ICU, TTE images were inadequate in over 60% of cases, while TEE provided new, clinically significant diagnoses in 28% that were not seen by TTE.7 The American College of Cardiology deems TEE a reasonable first test (versus TTE) in patients who are intubated, recently postoperative, or who have chronic obstructive pulmonary disease (COPD) or chest wall abnormalities, which may impair TTE examinations.8

In addition, there are certain structures and pathologies that are best visualized using TEE. These include the mitral valve, prosthetic valves, great vessels, and atria, as well as endocarditis and intracardiac thrombi (Table 22–3).8,9 This is particularly important in trauma patients who may be at risk of cardiac tamponade, aortic injuries, and traumatic valvular disruption or septal defects. Finally, TEE is indicated in critically ill patients requiring echocardiographic evaluation in whom TTE images are inadequate.

As with TEE in the general population, transesophageal echocardiography has a very low incidence of complications in critical care patients.2,10 Excluding feeding tube dislodgement, TEE had a complication rate of 2.6% in a review comprising 2508 examinations in the ICU.2 The most common complications included oropharyngeal mucosal lesions (0.7%), hypotension (0.6%), and coughing (0.3%). TEE should not be performed in any patient with esophageal perforation or in whom esophageal pathology or other comorbidities prevent safe and atraumatic insertion of the TEE probe.

Patient Preparation

Prior to performing a TEE examination, appropriate equipment for airway management, suctioning, and resuscitation should be available. Standard monitoring should be used including pulse oximetry, invasive or noninvasive blood pressure monitoring, and continuous electrocardiogram. It may be beneficial to have a second physician present to assist with management during the examination of hemodynamically unstable patients.

For elective TEE examinations, the patient should fast for 4 hours prior to the procedure to facilitate gastric emptying. Topical anesthesia with aerosolized or viscous local anesthetic should be performed prior to probe insertion. Judicious sedation may be used at the intensivist's discretion, depending on the patient's clinical presentation. Small titrated doses of benzodiazepines and/or opioids are commonly used, with reversal agents readily available.

For emergent TEE, gastric contents should first be suctioned via orogastric or nasogastric tube if in situ. In a patient with cardiopulmonary instability and an unsecured airway, endotracheal intubation prior to TEE may be clinically indicated. While TEE can be an invaluable tool for diagnosis, its use should not delay appropriate resuscitation and management of the unstable patient.

Imaging

Many TEE studies in the ICU are performed to facilitate diagnosis and management of an unstable patient, or to rule out a specific disease such as tamponade or endocarditis. Thus, while a comprehensive TEE exam as described in the ASE/SCA guidelines should be performed in all cases,11 the intensivist performing echocardiography should initially focus the exam based upon the patient's clinical status and suspected diagnosis. The following is a brief summary of the intensivist's echocardiographic approach to various clinical presentations and diagnoses—a thorough discussion of each topic can be found in earlier chapters.

Hypovolemia and Preload Assessment

Significant hypovolemia is seen on TEE as decreased left ventricular (LV) filling and commonly demonstrated by obliteration of the LV cavity (“kissing” papillary muscles) in the transgastric (TG) midpapillary short-axis (SAX) view. However, systolic cavity obliteration may also occur in the setting of reduced afterload and/or increased ejection fraction in 20% of cases, and may not reflect hypovolemia.12 Furthermore, visual detection of small (< 10%) yet potentially important changes is difficult and more objective measurements of LV volumes are required, especially in patient with compromised ventricular function (see Chapter 21). A left ventricular end-diastolic area (LVEDA) in the TG SAX view less than 6.3 cm2/m2 (women) or 7.5 cm2/m2 (men) could be consistent with hypovolemia as these values define the lower limits of normal in subjects under general anesthesia.13 Unfortunately, there is a wide range of normal values and as a consequence, visual estimation of LV volume is limited to the extremes—cavity obliteration or marked ventricular dilation.

Despite the fact that pulmonary artery occlusion pressure (PAOP) does not accurately reflect preload or volume responsiveness, it is still used as supportive criterion for the diagnosis of acute respiratory distress syndrome and heart failure. Noninvasive estimation of PAOP is feasible using TEE-derived simple Doppler variables, but not in every patient. In fact, Vignon and colleagues14 prospectively assessed the ability of TEE to predict PAOP in mechanically ventilated patients with a pulmonary artery catheter. In the initial group of patients, the correlations between simple Doppler variables and PAOP were better in patients with depressed LV systolic function than in those with normal LV systolic function. PAOP 18 mm Hg or less could be predicted by an E-wave/A-wave ratio greater than 1.4, E-wave deceleration time greater than 100 milliseconds, atrial filling fraction greater than 31%, and systolic fraction of pulmonary venous flow greater than 44%. In a second group, these cutoff values were prospectively validated and the utility of color M-mode Doppler flow propagation velocity (Vp) and the maximal early diastolic velocity of the lateral mitral annulus by tissue Doppler (E′) were assessed. An E/E′ ratio less than 8 and an E/Vp ratio less than 1.7 were predictive for PAOP 18 mm Hg or less, but the predictive value of these variables was not different from the “simpler” pulsed Doppler variables.14

Regional Wall Motion Abnormalities

A thorough survey of the left ventricle using the midesophageal (ME) four-chamber, two-chamber, and long-axis (LAX) views, as well as the TG basal, midpapillary, and apical SAX views allows the intensivist to identify regional wall motion abnormalities. It is important to examine each segment for thickening as well as motion, and to observe each segment in multiple views. Comparison with previous echocardiographic exams is helpful to determine whether any abnormalities are new or pre-existing. A new regional wall motion abnormality is consistent with myocardial ischemia. Although the sensitivity for ischemia detection by TEE is established, the specificity and implications of a transient wall motion abnormality are less clear (see Chapter 21).

Left Ventricular Function

Contractility can be inferred from the left ventricular ejection fraction (LVEF). There are several methods by which TEE can estimate the LVEF, including fractional area change, visual assessment, Simpson's rule, area-length method, and three-dimensional (3D) modeling. Cardiac output can be calculated using the continuity equation from the LV outflow tract area and velocity-time integral. These techniques are described in earlier chapters. It is also important to note the presence of associated left ventricular hypertrophy or dilatation.

Right Ventricular Function

Although right ventricular function can be more difficult to quantify than left ventricular function, it is possible nonetheless to identify patients with impaired RV performance. Acute right ventricular dysfunction in the critically ill patient may signify pulmonary embolus, acute pulmonary hypertension, or right coronary artery disease. The right ventricle is a thin-walled structure; thus, RV thickening should not be used to quantify function. Tricuspid annular displacement toward the RV apex in the ME four-chamber view, fractional area change in the TG short-axis view, and 3D imaging can all be used to quantify RV function. Other signs of RV dysfunction include flattening of the septum (creating a “D-shaped” left ventricle) and right ventricular enlargement.

Diastolic Function

Impaired cardiac filling can cause cardiac failure in the setting of normal systolic function. Diastolic dysfunction differs from systolic dysfunction in pathophysiology and therapeutic approach; therefore, this diagnosis is important for the intensivist in planning management. A clinical finding of acute cardiogenic pulmonary edema, in the presence of normal systolic function by TEE, is highly suggestive of diastolic dysfunction.15 Diagnosis of diastolic dysfunction with TEE requires the use of pulsed-wave Doppler of mitral inflow and pulmonary venous flow (see Chapter 12). However, Sadler et al16 confirmed the load dependence of the pulsed-wave Doppler parameters used to evaluate LV diastolic properties. The abrupt preload reduction induced by ultrafiltration in ICU patients with renal failure significantly decreased E-wave maximal velocities and the E/A ratio because A-wave velocities remained unaffected. Both the isovolumic relaxation time and E-wave deceleration time were significantly prolonged by volume reduction. Also, changes in the pulmonary vein D wave paralleled that of mitral E wave, and the S/D ratio increased after hemodialysis. Therefore, conventional pulsed-wave Doppler parameters to identify transient LV diastolic dysfunction in clinical settings characterized by abrupt changes in loading conditions should be interpreted with caution.17 On the other hand, E′ velocity from the lateral mitral annulus may be less preload dependent.17

Valvular Function

Each valve should be examined for stenosis and regurgitation, as well as abnormalities such as vegetations, thickening, calcification, and prolapse. While valvular lesions may cause hemodynamic instability, their presence may also have implications for management. Severe tricuspid regurgitation, for instance, may result in spurious measurements of cardiac output using thermodilution. Mitral regurgitation may occur as a result of ischemic disease and papillary muscle rupture. Aortic stenosis may be associated with secondary diastolic dysfunction as a result of ventricular remodeling.

Cardiac Tamponade

In the intensive care unit, cardiac tamponade may occur following cardiac surgery, traumatic injury, pericardial effusion, or as a result of myocardial infarction with wall rupture. Rarely, tamponade may occur as a complication of central venous line placement or other procedures. Clinically, cardiac tamponade may present with Beck's triad of hypotension, elevated jugular venous pressure, and muffled heart sounds. However, the absence of these symptoms does not exclude the diagnosis of tamponade, as lesser volumes or chronic accumulation may present with more subtle clinical findings. TEE is highly sensitive and specific in the detection of pericardial tamponade, with accuracy greater than 99%.9 Pericardial fluid can often be detected in the TG midpapillary SAX, TG LAX, and ME four-chamber views (Figure 22–1). Restriction of cardiac filling and equalization of pressures between the left and right heart may manifest via flattening of the interventricular septum, decreased left ventricular end-diastolic volume, decreased LV size during inspiration, and diastolic posterior motion of the RV.

Miscellaneous

Other causes of unexplained hypotension that may be diagnosed with TEE include pulmonary embolus, hypertrophic obstructive cardiomyopathy, and intracardiac thrombus.

Pulmonary embolus is not normally seen with TEE, with the exception of central emboli involving the pulmonary trunk or main pulmonary arteries. Echocardiographic signs suggestive of pulmonary embolus include new-onset RV dilatation and dysfunction, acute tricuspid regurgitation, and elevated right-side pressures, in the appropriate clinical context.

Hypertrophic cardiomyopathy may result in dynamic obstruction of the left ventricular outflow tract (LVOT) in up to 25% of patients. This obstruction may be intermittent, and is typically exacerbated by hypovolemia, tachycardia, increased contractility, and decreased afterload. In the ME four-chamber and ME aortic valve (AV) LAX views, left ventricular hypertrophy is often seen, particularly septal hypertrophy. Diastolic dysfunction is common in this population. Left ventricular outflow obstruction can cause high flow velocities in the LVOT, increased LV pressures, and a dagger-shaped pattern of flow on pulsed-wave Doppler or continuous-wave Doppler of the outflow tract.

Intracardiac thrombi can be seen in patients with ventricular aneurysm, dilated cardiomyopathy, coagulopathic states, and dysrhythmias. Thrombi appear as echogenic, mobile masses that may be freely floating or attached to a cardiac structure, and may mimic intracardiac tumors. Clinically, pulmonary or systemic thromboembolic disease may arise as a result.

TEE can be a useful diagnostic tool in the critically ill patient with unexplained hypoxemia. Causes of hypoxemia that can be detected with TEE include left ventricular failure, valvular pathology, cor pulmonale, pulmonary hypertension, pulmonary vein stenosis, central pulmonary embolus, right to left shunts, and pleural effusions (Figure 22–2).18

Echocardiographic findings consistent with pulmonary hypertension include right ventricular dilatation and dysfunction. In the presence of tricuspid regurgitation, the pulmonary artery systolic pressure (PASP) can be estimated by the modified Bernoulli equation (PASP = 4(VTR)2 + CVP, where VTR is the maximal flow in m/s of the tricuspid regurgitant jet and CVP is central venous pressure). Pulmonary vein stenosis is an uncommon cause of pulmonary hypertension whose incidence has been increasing as a result of transcatheter radiofrequency pulmonary vein isolation for the treatment of atrial fibrillation.

Right-to-left shunts may also cause unexplained hypoxemia in the ICU in patients with elevated right heart pressures. Shunts may be intracardiac (patent foramen ovale, atrial septal defect, ventricular septal defect) or intrapulmonary. Intracardiac shunting can be detected with color-flow Doppler, or by the venous injection of agitated saline for contrast.

Febrile patients in the ICU may require evaluation for the presence of endocarditis. In these patients, TEE is the gold standard for noninvasive assessment of endocarditis, with a sensitivity of 88% to 98%.19–21 In contrast, TTE has a sensitivity of 58% to 62%.19,21 TEE should be performed to rule out endocarditis in patients with a high index of suspicion and a negative TTE, for suspected prosthetic valve endocarditis, in patients with bacteremia of unknown source (or refractory to antibiotic therapy), and to assess complications of known endocarditis.22

Echocardiographic signs of infective endocarditis include vegetations on valves, cardiac structures, or devices; dehiscence of prosthetic valves; new onset of valvular regurgitation; and abscesses.23 Vegetations on native or prosthetic valves occur most commonly on the atrial side of the mitral and tricuspid valves and the ventricular side of the aortic and pulmonary valves (Figure 22–3). Vegetations are frequently oscillatory, moving in sync with the cardiac cycle and valvular motion. Repeat TEE examinations are often necessary over the course of treatment to assess for complications of infective endocarditis and the disappearance of vegetations.

In the patient with unexplained stroke or peripheral embolic disease, TEE may be helpful to identify intracardiac sources of emboli. These include left-side intracardiac thrombi, endocarditis, tumors, aortic atheromatous disease, and right-side masses in the presence of a right-to-left shunt. Patients with indwelling peripherally inserted central catheter (PICC) lines are at increased risk of catheter-associated thrombus (Figure 22–4), occurring in 15% of cases in a recent prospective study.24 Prior to elective cardioversion of patients with atrial fibrillation or flutter, TEE may be indicated to rule out the presence of thrombi in the left atrium, left atrial appendage, and left ventricle.20 Patients with a history of atrial dysrhythmia greater than 48 hours who are not therapeutically anticoagulated are at increased risk of thrombus development and stroke following cardioversion.

TEE is of particular value in the unstable trauma patient, particularly those with blunt or penetrating chest trauma. In such patients, TTE is most often unsatisfactory due to contusions, wounds, subcutaneous emphysema, mechanical ventilation, and/or difficulty with positioning. Patients with traumatic chest injury are at risk for aortic dissection and injury, myocardial contusion and acute dysfunction, traumatic ventricular septal defect, cardiac tamponade, and hemothorax, all of which can be visualized by TEE. TEE should not be performed if esophageal perforation is suspected.

Aortic Dissection

While angiography remains the gold standard for diagnosis of aortic injury in the trauma patient, it is time-consuming and requires transportation of a patient who may be hemodynamically unstable. TEE can be performed safely and rapidly at the bedside or in the operating room.25 The diagnostic accuracy of TEE in patients with suspected acute traumatic aortic injury is comparable to helical computed tomography (CT) but TEE was more sensitive for injuries involving the intimal or medial layers of the thoracic aorta.26 TEE may miss lesions in the distal ascending aorta and proximal aortic arch due to the interposition of the air-filled trachea between the aorta and the TEE probe.

In a patient with suspected aortic injury, careful inspection of the ascending aorta, aortic arch, and thoracic aorta should be performed. Echocardiographic signs of aortic dissection or disruption include presence of an intimal flap (Figure 22–5), turbulent blood flow in the aorta, evidence of low flow in the aorta visible as spontaneous echo contrast or “smoke,” aortic aneurysm, and thrombus.25

TEE is a minimally invasive diagnostic tool that can provide valuable information rapidly and can be performed at the patient's bedside. TEE is increasingly being used by critical care physicians, particularly in hemodynamically unstable and trauma patients. The information obtained from a TEE examination by a skilled operator will often lead to important management changes, as well as alert the intensivist to new diagnoses. Lastly, it is likely that the advent of real-time 3D imaging will only extend the usability of TEE for intensive care physicians.