General Indications

Ultrasonography has become an invaluable tool in the management of critically ill patients. Its safety and portability allow for use at the bedside to provide rapid, detailed information regarding the cardiovascular system¹ and the function and anatomy of certain internal organs. It also can be used by the clinician to assess the pleural and intraabdominal spaces and to perform some invasive procedures safely. General indications for performance of echocardiography in the ICU are listed in Box W2-1. Box W2-2 lists major indications for performance of primary TEE in the ICU. Other indications for use of bedside ultrasonography by the intensivist in critically ill patients are listed in Box W2-3.

Technical Aspects

Positioning

Proper positioning of the patient is important for obtaining an adequate image. For performance of TTE and TEE in patients who are not mechanically ventilated, optimal imaging usually is obtained by having the patient in the left lateral decubitus position. Taking the extra 5 minutes to position the patient on their left side for TTE often results in much improved image quality and minimizes aspiration risk for TEE. Adequate positioning of the patient varies depending on the structures being assessed for other ultrasound imaging (e.g., pleural space, peritoneal cavity, vascular structures, or bladder). Care must be taken when positioning a critically ill patient in bed, because these patients often have multiple vascular catheters, an endotracheal tube, drains, and other tubes or devices connected to them. When the ultrasound examination is done to localize and mark pleural or abdominal fluid collections for subsequent drainage, it is crucial that the patient remain in the same position used during the marking procedure until the actual drainage of the collection is performed. Risks of perforating surrounding organs (e.g., heart, spleen, liver, lungs, or bowel) and inducing significant morbidity are increased if the drainage is performed in a position different from the one used during marking.

Sedation

To optimize the ultrasound examination, the patient must be cooperative and nonagitated. Noninvasive procedures such as TTE and abdominal ultrasound usually are well tolerated by patients, and additional sedation rarely is needed to perform these procedures. When performing TEE, however, certain precautions need to be taken. Patients should fast (or have their tube feeds stopped) for at least 4 hours before the procedure. Topical anesthesia of the oropharynx also is helpful before insertion of the TEE probe, especially in patients who are not endotracheally intubated.³ Even if adequate topical anesthesia is provided, insertion of the TEE probe still can cause significant discomfort and anxiety, so providing adequate sedation and analgesia is important. Frequently used sedative or analgesic agents include intravenous (IV) midazolam, fentanyl, and propofol. Dosing should be titrated according to clinical parameters including arterial blood pressure, minute ventilation, and arterial oxygen saturation.³ Sedative-induced hypotension is a frequent problem in patients with depressed ventricular function or decreased systemic vascular resistance, and occasionally patients may require transient support with IV volume infusion or rarely a vasopressor agent. If the patient is extremely uncooperative and biting, transient paralysis accompanied by increased sedation may have to be used to perform TEE safely.

Transthoracic Versus Transesophageal Echocardiography in a Critically Ill Patient

Accurate and prompt diagnosis is crucial in the ICU. The easiest and least invasive way to image cardiac structures is TTE.³ This noninvasive imaging modality is of great value in the critical care setting because of its portability, widespread availability, and rapid diagnostic capability. In the ICU, TTE in certain cases may fail to provide adequate image quality because of different factors that potentially can hinder the quality of the ultrasound signal, as was described previously. The failure rate (partial or complete) of TTE in the ICU has been reported to be 30% to 40%.^10,11 Improvements have been made in transthoracic imaging (e.g., harmonics and contrast and digital technologies), however, resulting in a lower failure rate of TTE in the ICU (10%-15% in our institution).

TEE is particularly useful for evaluation of suspected aortic dissection, prosthetic heart valves (especially in the mitral position), source of cardiac emboli, valvular vegetations, possible intracardiac shunts, and unexplained hypotension. TEE allows better visualization of the heart in general and especially the posterior structures, owing to the proximity of the probe and favorable acoustic transmission.¹ TTE also has limitations, however. For several areas of the heart and great vessels, TEE may provide limited images. The view of the left ventricular apex often is foreshortened with TEE, and an apical left ventricular clot can be missed. TTE usually is superior for visualization of the apex. Because of interposition of the left mainstem bronchus, the superior portion of the ascending aorta is another important area that may not be well visualized with TEE. With TEE, transducer position and angulation are constrained by the relative positions of the esophagus and heart. The relatively fixed relationship between the position of the probe and the heart often makes it impossible to align the Doppler beam parallel to the flow of interest (e.g., to evaluate the jet of blood resulting from aortic stenosis). In addition, the 2D image planes of TEE often make standard anatomic measurements more difficult to obtain.

As a result of the significantly improved technical quality of TTE, most ICU patients can be studied satisfactorily with this modality. Immediate TEE is still preferable, however, in certain specific clinical situations in which TTE is likely to fail or be suboptimal.¹¹ The major indications for primary TEE in the ICU^12,13 are listed in Box W2-2. Even when TEE is necessary, data from the TTE examination are often essential for the final clinical interpretation.

Hemodynamic Evaluation

Ventricular Function

Left Ventricular Systolic Function

Evaluation of left ventricular performance by echocardiography is often paramount in the ICU. Accurate and timely assessment of systolic function should be an integral part of the medical management of hemodynamically unstable critically ill patients. Global assessment of left ventricular contractility includes the determination of ejection fraction (EF), circumferential fiber shortening, and cardiac output.

The simplest quantitative approach is to measure the mid-left ventricular short-axis dimension at end diastole and end systole for determination of the percent fractional shortening. Fractional shortening is related directly to EF; normal fractional shortening is 30% to 42%.¹

In the setting of regional wall motion abnormalities, fractional shortening may underestimate or overestimate global ventricular function and must be interpreted in light of what is seen in all of the 2D imaging planes of the ventricle.¹⁴

Global systolic ventricular function also can be assessed quantitatively by fractional area change (normal value is 36% to 64%)¹⁵ and EF (normal value is 55% to 75%) (Figure W2-1):

Figure W2-1 Fractional area change and ejection fraction calculation. Endocardial contour of the left ventricular cavity is traced at end diastole (A) and at end systole (B) in the transthoracic apical four-chamber view. Machine-integrated software computes the data and gives corresponding end-diastolic and end-systolic areas. Fractional area change can be calculated with these data (C). Normal values are 36% to 64%.¹⁵ Corresponding end-diastolic (D) and end-systolic (E) volumes are computed using the modified Simpson’s method. The data are used to calculate the ejection fraction (F). Normal values are 55% to 75%. LA, left atrium; LV, left ventricle; RA, right atrium; RV, right ventricle.

These measurements require good image quality, because endocardial border contours must be traced (see Figure W2-1). Machine-integrated software computes the data and provides volumes, areas, and the resultant EF (see Figure W2-1). In patients with regional wall motion abnormalities, more precise measures of stroke volume can be made by approximating ventricular volumes as a stack of elliptical discs on biplane imaging (modified Simpson’s method).^1,15

In the critical care setting, endocardial border definition may be suboptimal because of poor image quality.^10,16,17 In these cases, global ventricular function often is assessed qualitatively by visual inspection alone. This method has been found to be reliable when used by experienced clinicians.¹⁸ By simple visualization of the kinetics and size of the cardiac cavities in real time, an experienced intensivist with a sufficient echocardiographic background can establish a functional diagnosis immediately.

Analysis of regional wall motion includes a numerical scoring system to describe the movement of the different regions of the left and right ventricle (1 = normokinesia; 2 = hypokinesia; 3 = akinesia; 4 = dyskinesia; 5 = aneurysmal change).¹⁵ Visualized from the short-axis view of the left ventricle, a complete overview of myocardial areas perfused by the three major coronary arteries can be obtained (Figure W2-2). If the TTE examination is technically difficult and the endocardium is poorly visualized, harmonic imaging and possibly contrast, if needed, can dramatically improve endocardial border visualization and subsequent evaluation of global systolic function (as discussed further later in this chapter). For the remaining few technically challenging cases with suboptimal TTE, performance of TEE allows for a more precise evaluation of ventricular function in most critically ill patients because of the higher image quality that can be obtained with this echographic modality.

Figure W2-2 Transthoracic short-axis echocardiographic view of the left (LV) and right (RV) ventricles at the midpapillary muscle level. In this tomographic view of the heart, areas of myocardium and papillary muscles (AL, anterolateral; PM, posteromedial) supplied by all three major coronary arteries are represented. ANT, anterior; AS, anteroseptal; INF, inferior; IS, inferoseptal; LAT, lateral.

Left Ventricular Failure in the Intensive Care Unit

In a critically ill patient with unexplained hemodynamic instability, determination of cardiac function is an integral part of the medical management. Echocardiography is valuable in this setting because the clinical examination and invasive hemodynamic monitoring often fail to provide an adequate assessment of ventricular function. In a study by Fontes et al.¹⁹ that compared pulmonary artery (Swan-Ganz) catheterization and TEE, the overall predictive probability for conventional clinical and hemodynamic assessment of normal ventricular function was 98%, whereas for abnormal ventricular function (EF <40%), it was 0%. Several other studies have reported similar results.^20–22 Assessment of biventricular function is one of the most important indications for performance of echocardiography in the ICU. In a study by Bruch et al.,²³ 115 critically ill patients were studied by TEE. The most common indication for TEE was hemodynamic instability (67% of patients). Of these hemodynamically unstable patients, 20 (26%) were found to have significant left ventricular dysfunction (EF <30%). In a study by McLean²⁴ of the use of TEE in the ICU, the most common reason to request a TEE was assessment of left ventricular function. In most patients, left ventricular function was assessed adequately by TTE before TEE. In a study by Vignon et al.,¹⁷ TTE allowed adequate evaluation of global left ventricular function in 77% of mechanically ventilated ICU patients. Although TEE was needed for most other indications, TTE was shown to be an excellent diagnostic tool for assessment of left ventricular function in the ICU (Figure W2-3) even when positive end-expiratory pressure is present.

Figure W2-3 Dilated cardiomyopathy. Transthoracic examination of a severely dilated left ventricle (LV) in the parasternal long-axis (A) and apical four-chamber (B) views. The 65-year-old patient presented with flash pulmonary edema and later was found to have severe diffuse coronary artery disease. LA, left atrium; RA, right atrium; RV, right ventricle.

Several important points should be emphasized: (1) Significant left ventricular dysfunction is common in critically ill patients; (2) ventricular function should be assessed in all patients with unexplained hemodynamic instability, because this information is particularly important for guiding resuscitation and informing decisions regarding subsequent medical or surgical management; (3) it is now possible to obtain adequate information about ventricular function in most ICU patients using TTE, but TEE provides better accuracy in patients with suboptimal imaging by TTE.

Sepsis-Related Cardiomyopathy

Classically, septic shock has been considered a “hyperdynamic” state characterized by normal or high cardiac output. Echocardiographic studies indicate that ventricular performance often is markedly impaired in patients with sepsis.^25–27 Parker et al.²⁸ were the first to describe left ventricular hypokinesis in septic shock. They reported that survivors manifested severely depressed left ventricular EF, but that adequate left ventricular stroke output was maintained as a result of acute left ventricular dilation.²⁹ Jardin et al.²⁵ studied 90 patients with septic shock and performed daily bedside assessments of left ventricular volume and left ventricular EF using TTE. They observed that left ventricular EF was significantly depressed in all patients, resulting in severe reductions in left ventricular stroke volume. Of these patients, 34 (38%) eventually were weaned from hemodynamic support and showed gradual improvement in left ventricular EF and ultimately recovered. The remaining 56 patients (62%) eventually died (of early circulatory failure or late multiple organ failure). In this subset, the degree of left ventricular dysfunction was less than in survivors but failed to improve over time. The severity of left ventricular dysfunction does not predict outcome. A paradoxical relationship between the degree of left ventricular dysfunction and the likelihood of recovery also has been described by others.^25,28,30,31 Among patients who survive, left ventricular dilation and systolic dysfunction usually are reversible.

Left ventricular EF might not be a reliable index of left ventricular systolic function in patients with early septic shock because this is a state characterized by low systemic vascular resistance that unloads the left ventricle.²⁵ Normal or supranormal EF in early sepsis might lead clinicians to make the wrong inference about cardiac reserve because left ventricular EF might decrease if afterload is increased by the administration of vasopressor agents.

Left Ventricular Diastolic Function

In the ICU, diastolic dysfunction should be suspected when ventricular filling pressure (pulmonary capillary wedge pressure) is elevated and EF is normal or supranormal.¹ The diastolic properties of the ventricle often are assessed by evaluating Doppler echocardiographic mitral inflow and pulmonary venous flow patterns. Mitral inflow, as measured by pulsed wave Doppler at the tips of the mitral leaflets, is characterized by an early filling phase (E wave) followed by atrial systole, resulting in additional filling (A wave) (Figure W2-4). The transmitral Doppler pattern always should be interpreted in conjunction with pulsed wave Doppler of the pulmonary venous flow, which is characterized by a systolic phase (S), a diastolic phase (D), and an atrial phase (AR) from reversal of flow into the pulmonary veins during atrial contraction (Figure W2-5). These filling patterns are related to the intrinsic diastolic properties of the myocardium and are influenced by many different factors, particularly left atrial pressure, heart rate, ischemia, ventricular hypertrophy, and valvular pathologies. Only modest correlation has been found between Doppler indices of diastolic function and parameters measured using more invasive means.^32,33 Integrated interpretation of mitral and pulmonary venous flow patterns may be useful for diagnosing abnormal myocardial relaxation (e.g., owing to hypertensive heart disease, hypertrophic cardiomyopathy, or coronary ischemia) or restrictive pathology (e.g., owing to cardiomyopathy, constrictive pericarditis, coronary artery disease, cardiac transplantation, or dilated cardiomyopathy). Nevertheless, these findings must be interpreted with caution when caring for critically ill patients, given the many different factors that can acutely influence flow patterns in this population of patients.

Figure W2-4 Normal mitral inflow profile as measured by transthoracic pulsed wave Doppler at the tips of the mitral leaflets. It is characterized by an early filling phase (E wave) followed by atrial systole (A wave), which results in additional filling. These filling parameters are related to intrinsic diastolic myocardial properties and can be influenced by many different factors (see text).

Figure W2-5 Normal pulmonary venous flow profile as measured by transthoracic pulsed wave Doppler with the sample volume placed in the right superior pulmonary vein. It is characterized by a predominant systolic wave (S), a diastolic wave (D), and an atrial wave (AR) (from reversal of flow into the pulmonary veins occurring during atrial contraction). The pulmonary venous flow profile always should be interpreted in conjunction with the transmitral Doppler pattern to have a more complete assessment of diastolic function.

Right Ventricular Function and Ventricular Interaction

Abnormal right ventricular function often plays an important and sometimes underestimated role in the pathogenesis of critical illness.^34–36 Based on an echocardiographic definition,³⁷ massive pulmonary embolism and acute respiratory distress syndrome are the two main causes of acute cor pulmonale in adults.³⁸ In the critical care setting, right ventricular function also can be altered by any other perturbations that increase right ventricular afterload, such as positive end-expiratory pressure or increased pulmonary vascular resistance (from vascular, cardiac, metabolic, or pulmonary causes). Depressed right ventricular systolic function is also often associated with right ventricular infarction, most commonly in the setting of inferior myocardial infarction. Acute sickle-cell crisis, air or fat embolism, myocardial contusion, and sepsis are other causes of acute right ventricular dysfunction.

Adequate assessment of right ventricular function is important when caring for hemodynamically unstable, critically ill patients, specifically patients with massive pulmonary embolism and acute respiratory distress syndrome, because the diagnosis of concomitant significant right ventricular dysfunction may alter therapy (e.g., fluid loading, use of vasopressors, use of thrombolytics) and provide information about prognosis.^38,39 Echocardiographic examination of the right ventricle requires primarily an assessment of the size and kinetics of the cavity and septum.^37,40 Normally the right ventricle appears relatively flat. As it dilates, the apical region of the right ventricle becomes more rounded (Figure W2-6). In the short-axis view, the right ventricle, which usually has a crescentic shape, becomes oval because of septal displacement and bulging of the right ventricular free wall (see Figure W2-6).¹ Right ventricular size and function generally are evaluated by visual comparison with the left ventricle. Right ventricular diastolic dimensions can be obtained by measuring right ventricular end-diastolic area in the long axis, from an apical four-chamber view, using either TTE or TEE.

Figure W2-6 Severe right ventricular failure and dilation. A, Normal transthoracic parasternal short-axis view of the left (LV) and right (RV) ventricles at the midpapillary muscle level. B, Normal transthoracic apical four-chamber view of the left ventricle and right ventricle. These pictures of a normal heart depict the relationship between the left ventricle and right ventricle, with the left ventricle being normally larger than the right ventricle and the interventricular septum bulging slightly toward the right ventricle. C, Transthoracic parasternal short-axis view of the left ventricle and right ventricle in a patient with severe right ventricular failure and dilation. The right ventricular cavity is seen to be much larger than the left ventricular cavity. Because of the high volume and pressure in the right ventricle, the interventricular septum is bulging toward the left. This gives the left ventricle a characteristic “D” appearance. D, Transthoracic apical four-chamber view of the same patient shows the inverse relationship between the left ventricular and right ventricular sizes. The right ventricular dilation can occur only if associated with a proportional reduction in left ventricular diastolic dimension (“ventricular interaction”). This reduction in left ventricular diastolic dimension significantly impairs left ventricular relaxation and changes the pressure-volume relationship of the left heart chambers. LA, left atrium; RA, right atrium.

Because pericardial constraint necessarily results in left ventricular restriction when the right ventricle acutely dilates (i.e., there is ventricular interaction), one of the best ways to quantify right ventricular dilation is to measure the ratio between the right ventricular and left ventricular end-diastolic areas, an approach that cancels out individual variations in cardiac size.^37,40 Moderate right ventricular dilation corresponds to a diastolic ventricular ratio greater than 0.6; severe right ventricular dilation corresponds to a ratio greater than or equal to 1.^37,40 Right ventricular diastolic enlargement usually is associated with right atrial dilation, inferior vena caval dilation, and tricuspid regurgitation. When pressure in the right atrium exceeds pressure in the left atrium, the foramen ovale may open. Pressure and volume overload of the right ventricle can lead to distortion of left ventricular geometry and abnormal motion of the interventricular septum. With conditions of high strain imposed on the right ventricle (volume or pressure overload or both), the interventricular septum flattens, and the left ventricle appears to have a “D” shape (see Figure W2-6).^4,37 This “paradoxical” septal motion also is seen at the interatrial level.

Because the two ventricles are enclosed within the relatively stiff pericardium, the sum of the diastolic ventricular dimensions has to remain constant.⁴¹ Acute right ventricular or left ventricular dilation can occur only if it is associated with an acute and proportional reduction in left ventricular or right ventricular diastolic dimension (i.e., ventricular interaction). With acute right ventricular dilation, septal displacement impairs left ventricular relaxation; the opposite occurs with acute left ventricular dilation. In these situations, the pressure-volume relationships of the left and right heart chambers are altered, and information obtained from a pulmonary artery catheter could be misleading (e.g., high filling pressures are recorded despite normal or even low circulating volume).

Assessment of Cardiac Output

Measurement of cardiac output remains a cornerstone in the hemodynamic assessment of critically ill patients. Thermodilution is considered the gold standard approach for determining cardiac output in most ICUs. Measurement of cardiac output using thermodilution requires placement of a pulmonary artery catheter (or at least central venous and arterial catheters); although a useful technique, it is invasive and potentially inaccurate. Unreliable values are particularly common in the presence of tricuspid regurgitation related to high pulmonary artery pressure. Several methods for determining cardiac output have been described using 2D and Doppler echocardiography. With this technique, stroke volume and cardiac output can be determined directly by combining Doppler-derived measurements of instantaneous blood flow velocity through a conduit with the cross-sectional area of the conduit. Blood flow can be calculated through various cardiac structures, including the pulmonary valve,⁵² mitral valve,^53,54 and aortic valve.^55–58 In the absence of intracardiac shunts, blood flow through these structures should be the same (continuity equation).⁵⁹ Of these methods, the one using the left ventricular outflow tract and aortic valve as the conduit is probably the most reliable and most commonly used. There is excellent agreement with thermodilution in most situations.^55–58 The left ventricular stroke volume is obtained by measuring the cross-sectional area of the left ventricular outflow tract (area [cm²] = (left ventricular outflow tract diameter [cm²]) × (π/4), assuming that just below the aortic annulus, the left ventricular outflow tract is circular) multiplied by the transaortic flow velocity time integral derived from a spectral Doppler tracing. The stroke volume obtained is multiplied by the heart rate to give the cardiac output: cardiac output = cross-sectional area × velocity time integral × heart rate (Figure W2-7).

Figure W2-7 Calculation of the stroke volume and cardiac output from the left ventricular outflow tract (LVOT). A, Left ventricular outflow tract diameter obtained from the transthoracic parasternal long-axis view, just below the insertion of the aortic valve leaflets. In this example, the left ventricular outflow tract diameter is 2.17 cm. B, Doppler interrogation (with pulsed wave Doppler) is performed from the apical view with the sample volume being placed in the left ventricular outflow tract, just below the aortic valve. C, Spectral Doppler tracing from the left ventricular outflow tract from which the transaortic flow velocity time integral (VTI) is derived. In this example, the VTI is 17.3 cm. D, The left ventricular stroke volume (SV) is obtained by measuring the cross-sectional area (CSA) of the left ventricular outflow tract (area [cm²] = left ventricular outflow tract diameter [cm²] × π/4) and multiplying by the transaortic VTI derived from the spectral Doppler tracing. The stroke volume obtained is multiplied by the heart rate to give the cardiac output (CO). LA, left atrium; LV, left ventricle; RV, right ventricle.

With TTE, the left ventricular outflow tract diameter usually is obtained from the parasternal long-axis view, just below the insertion of the aortic valve leaflets. The Doppler interrogation is performed through the aortic valve from the apical view (see Figure W2-7). With TEE, the left ventricular outflow tract diameter usually is obtained from the five-chamber view of the left ventricle. The transgastric view usually is used to obtain an apical long-axis view of the aortic valve through which Doppler interrogation is performed.⁶⁰ With either TTE or TEE, obtaining an accurate left ventricular outflow tract diameter and Doppler signal is essential to have an accurate cardiac output calculation. Because the measure of the left ventricular outflow tract diameter has a second-order relationship with the cross-sectional area (see previous formula), it is crucial that this measure be determined precisely. For the Doppler signal to be reliable, the Doppler sample must be parallel to the transaortic flow with an angle of incidence not exceeding 20 degrees to avoid underestimation of transaortic velocity. Using TTE, McLean and coworkers⁶¹ showed an excellent correlation (r = 0.94) between cardiac output determined by the left ventricular outflow tract Doppler method and the thermodilution method in critically ill patients. Other studies have shown similar results.⁵⁵ In a study by Feinberg et al.,⁵⁸ cardiac output determined by TEE Doppler imaging was obtainable in 88% of 33 critically ill patients, and there was good correlation (r = 0.91) with the thermodilution method. Descorps-Declere et al.⁶⁰ also showed transgastric pulsed Doppler measurement across the left ventricular outflow tract with TEE to be a clinically acceptable method for cardiac output measurement in critically ill patients (r = 0.975 compared with the thermodilution method).

Another promising ultrasound-based technology to estimate cardiac output noninvasively in adults uses a small transesophageal Doppler probe to measure blood flow velocity waveforms in the descending aorta combined with a nomogram (based on height, weight, and age) for estimation of aortic cross-sectional area. This minimally invasive esophageal probe can be inserted easily in sedated patients and left in place safely for several days to provide continuous monitoring of cardiac function.^62,63 Several technical problems can limit the accuracy of cardiac output measurements by esophageal Doppler monitoring,⁶² however, and although initial results are promising,^64–66 more studies are needed to make a decision regarding the accuracy of this technique in critically ill patients.

Assessment of Filling Pressures and Volume Status

Adequate determination of preload and volume status is important for proper management of critically ill patients. Invasive pressure measurements to assess left ventricular filling are commonly used at the bedside to make inferences regarding left ventricular preload. These pressure measurements correlate only weakly with left ventricular volume, however.⁶⁷ Data from invasive monitoring using pulmonary artery catheterization may be misleading because ventricular compliance is altered secondary to numerous factors.^68,69 Differences in diastolic compliance among patients may account for the weak correlation between pressure and volume and may limit the ability to use pressure measurements alone to derive information concerning left ventricular preload.¹⁴ Echocardiography can be helpful for adequately assessing preload. Parameters that can be measured using 2D imaging are left ventricular end-diastolic volume and left ventricular end-diastolic area. Using Doppler interrogation, additional information—mainly transmitral diastolic filling pattern and pulmonary venous flow—can be obtained.

Two-Dimensional Imaging

Echocardiography has been validated for left ventricular volume measurements.¹⁵ Subjective assessment of left ventricular volume by estimating the size of the left ventricular cavity in the short-axis and long-axis views is often adequate to guide fluid volume therapy at the extreme ends of cardiac filling and function. More precise quantitative values are desirable, however, and can be obtained by using endocardial border tracing (as described earlier). The normal left ventricular end-diastolic volume as determined by echocardiography is 80 to 130 mL,¹⁵ and the normal left ventricular end-diastolic volume index is 55 to 65 mL/m².¹⁵ Left ventricular end-diastolic area measured in the left parasternal short-axis view at the level of the midpapillary muscle is commonly used to estimate volume status (Figure W2-8). The normal values for left ventricular end-diastolic area in the short-axis view are 9.5 to 22 cm².¹⁵

Figure W2-8 Calculation of left ventricular end-diastolic area in the transthoracic short-axis view at the level of the midpapillary muscle by endocardial contour tracing. Values of normal left ventricular end-diastolic area in the short axis range from 9.5 to 22 cm².¹⁵ The level of the midpapillary muscle is used because of the reproducibility of the view and because changes in left ventricular volume affect the short axis of the ventricle to a greater degree than the long axis. LV, left ventricle; RV, right ventricle.

Two-dimensional TTE evaluation of ventricular dimensions has been found to be useful in assessing preload and optimizing therapy of ICU patients.^25,70 Nevertheless, image quality may be suboptimal and preclude adequate visualization of the endocardial border by TTE. This potential limitation of TTE has been partly circumvented in recent years with the advent of harmonic imaging and contrast echocardiography (see later). In cases in which endocardial border visualization remains suboptimal, TEE is the modality of choice. With TEE, left ventricular volume can be estimated rapidly by subjective assessment of the left ventricular size. Quantitatively, it is estimated most often by determining left ventricular cross-sectional area at the end of diastole, most commonly using the transgastric short-axis view at the level of the midpapillary muscle. This section is used because of the reproducibility of the view and because changes in left ventricular volume affect the short axis of the ventricle to a greater degree than the long axis.¹⁴ The end-diastolic area must be measured consistently from the same reference section. End-diastolic area measured with TEE correlates with left ventricular volume determined by radionuclide studies.⁷⁰

Systolic obliteration of left ventricular cross-sectional area accompanies decreased end-diastolic area and is considered to be a sign of severe hypovolemia (Figure W2-9). Although a small end-diastolic area generally indicates hypovolemia, a large end-diastolic area does not indicate adequate preload in patients with left ventricular dysfunction. Also, when systemic vascular resistance is low, as in early sepsis, left ventricular emptying is improved because of the lowered afterload. In these situations, it may be difficult to differentiate hypovolemia from low systemic vascular resistance by echocardiography alone, because both conditions are associated with decreased end-diastolic area.

Figure W2-9 Systolic obliteration of the left ventricle (LV) in a patient with severe left ventricular hypertrophy and dehydration. This transthoracic parasternal short-axis view shows the left ventricle at end diastole (A) and at end systole (B). Nearly complete obliteration of the left ventricular cavity is seen at end systole. Systolic obliteration of the cross-sectional area accompanies decreased end-diastolic area and is considered to be a sign of severe hypovolemia. In this case, the patient presented with hypotension and was found to be severely dehydrated because of a viral gastroenteritis. RV, right ventricle.

Knowledge of left ventricular end-diastolic volume or absolute preload does not allow for accurate prediction of the hemodynamic response to alterations in preload.⁷¹ Tousignant et al.⁷² investigated the relationship between left ventricular stroke volume and left ventricular end-diastolic area in a cohort of ICU patients and found only a modest correlation (r = 0.60) between single-point estimates of left ventricular end-diastolic area and responses to fluid loading. Based on the assumption that changes in end-diastolic area occur because of changes in left ventricular volume, the determination of this area and its subsequent degree of variation after a fluid challenge could help better assess preload responsiveness. Studies have shown that changes in end-diastolic area measured by TEE using endocardial border tracing are closely related to changes in cardiac output and are superior to measurements of pulmonary artery occlusion pressure for predicting the ventricular preload associated with maximal cardiac output.⁷³

Circulating volume status also can be assessed by 2D echocardiography by indirectly estimating right atrial pressure; this is often done by assessing the diameter and change in caliber with inspiration of the inferior vena cava (Figure W2-10). This method has been shown to discriminate reliably between right atrial pressures less than 10 mm Hg or greater than 10 mm Hg.⁷⁴ A dilated vena cava (diameter >20 mm) without a normal inspiratory decrease in caliber (>50% with gentle sniffing) usually indicates elevated right atrial pressure. In mechanically ventilated patients, this measure is less specific because of a high prevalence of inferior vena cava dilation.^75,76 A small vena cava reliably excludes the presence of elevated right atrial pressure in these patients.^75,76

Figure W2-10 Indirect assessment of circulating volume status on two-dimensional echocardiography by assessing the diameter and change in caliber with inspiration of the inferior vena cava (IVC). This method has been shown to discriminate reliably between right atrial pressures of less than or greater than 10 mm Hg. A dilated vena cava (>20 mm) without the normal inspiratory decrease in caliber (>50% on gentle sniffing) usually indicates elevated right atrial pressure. A small vena cava reliably excludes elevated right atrial pressure in these patients. In this case, the IVC was dilated at 2.5 cm with minimal respiratory variation in a patient spontaneously breathing. The right atrial pressure was estimated to be approximately 10 to 15 mm Hg. Images were obtained in the subcostal view. HV, hepatic veins; RA, right atrium.

Doppler Flow Patterns

Information obtained by analysis of the Doppler signal at the level of the mitral valve and pulmonary vein offers additional information about preload.^77,78 These Doppler profiles can be obtained by either TTE or TEE. Transmitral parameters that have been studied include the relation of early to late transmitral diastolic filling (E/A ratio), isovolumetric relaxation time, and the rate of deceleration of early diastolic inflow (deceleration time).¹

A decrease in preload causes a significant reduction in the E wave (early filling flow wave) velocity at the mitral level in conjunction with a decrease of the S wave (systolic flow wave) in the pulmonary vein. In clinical practice, the E/A ratio is easy to assess; the normal value of this ratio is approximately 1.^1,3 In conjunction with normal left ventricular contractility, a low E/A ratio is usually a characteristic sign of inadequate preload.⁷⁹

Pulmonary venous flow also can be used to assess left atrial pressure. A normal pulmonary venous flow pattern showing a predominance of flow during systole (S phase) compared with early diastole (D phase) usually indicates that left atrial pressure is less than 8 mm Hg, whereas the opposite predominance of flow (in the absence of significant mitral regurgitation) usually indicates elevation of left atrial pressure.¹

Transmitral and pulmonary vein Doppler patterns strongly depend on intrinsic and external factors and are not affected purely by the loading conditions of the left ventricle. It is crucial that interpretation of Doppler parameters be done in conjunction with a global analysis of cardiac function and other available hemodynamic or anatomic variables.

Hypovolemia in the Intensive Care Unit

Precise and rapid assessment of volume status is crucial when caring for hemodynamically unstable ICU patients. Hypovolemia is one of the most common causes of hypotension in the ICU. As was discussed in detail earlier, bedside echocardiography offers a quick and reliable way of estimating volume status by evaluating cardiac dynamics and left ventricular dimensions and area. The finding of end-systolic cavity obliteration is usually a reliable sign of hypovolemia. Other changes in the volume status are usually associated with subtle changes in left ventricular cavity size, so only this extreme is reliable to make the diagnosis of hypovolemia by echocardiography. In general, TTE has good sensitivity for diagnosing the presence of a small hyperdynamic left ventricle, the most typical finding in hypovolemic patients with underlying normal cardiac function, although TEE is useful in the immediate postoperative setting (Video W2-1).

When dynamic left ventricular obstruction is present, cardiac output is low, and even in the presence of marked hypovolemia, pulmonary artery occlusion pressure is high. Paradoxical worsening of hypotension after intravascular volume loading may be the first clue to dynamic left ventricular obstruction in critically ill patients. It is important that this entity be recognized early and that the pathophysiologic process be well understood, because inadequate management of this condition can lead rapidly to worsening of hemodynamic status and death. Dynamic obstruction of the left ventricle can present in different forms. One of these forms is dynamic left ventricular outflow tract obstruction. Although dynamic left ventricular outflow tract obstruction is often seen in association with asymmetrical septal hypertrophy, it also can occur in other situations.^80,81 Dynamic left ventricular outflow tract obstruction is thought to be caused by the Venturi effect. This effect results when excessive acceleration of blood through a conduit produces a decrease in pressure. In the left ventricular outflow tract, such a decrease in pressure leads to a suction phenomenon that draws the anterior mitral leaflet and chordae inward toward the interventricular septum.⁸² This systolic anterior motion of the mitral valve leads to contact between the mitral leaflet and the septum that creates an obstructive subaortic pressure gradient and distortion of the mitral valve leaflet coaptation (Figure W2-11).⁸² By 2D echocardiography, the left ventricle appears to be small and hyperdynamic, and there is motion of the anterior leaflet (or chordae or both) toward the septum in systole (see Figure W2-11). With color Doppler, a “mosaic” pattern of flow is seen in the left ventricular outflow tract, owing to the high velocity and turbulence. Variable degrees of asymmetrical mitral regurgitation also may be present (see Figure W2-11). Continuous-wave Doppler shows the presence of a significant gradient in the left ventricular outflow tract. Dynamic left ventricular obstruction also can be present without systolic anterior motion. In the presence of reduced afterload, dehydration, or significant catecholaminergic stimulation, patients with a small hypertrophied left ventricle (typically seen in elderly patients with chronic hypertension) can develop midventricular obstruction due to hyperdynamic systolic obliteration of the left ventricular cavity (see Figure W2-9).⁸³ These physiologic factors may predict development or worsening of left ventricular dynamic obstruction. Interplay of these factors with preexisting ventricular hypertrophy predisposes the patient to develop cardiogenic shock from this combined loss of preload and presence of dynamic left ventricular obstruction. Dynamic left ventricular obstruction has also been described in patients with acute myocardial infarction, mostly in association with apical infarction.^81,84,85

Figure W2-11 Systolic anterior motion (SAM) of the mitral valve in a patient with asymmetrical left ventricular hypertrophy and dehydration. Two-dimensional transthoracic apical long-axis view shows movement of the anterior leaflet of the mitral valve (arrow) toward the interventricular septum during systole (A). This creates a subaortic dynamic obstruction. The resulting high velocity and turbulence in the left ventricular outflow tract (LVOT) gives a “mosaic” pattern of flow on color Doppler (B). A variable degree of asymmetric mitral regurgitation (MR) also may be present secondary to the systolic anterior motion, as shown in this example. LA, left atrium; LV, left ventricle; RV, right ventricle. (See Color Section in this text.)

In a study by Chenzbraun et al.⁸⁵ in ICU patients, four patients with hemodynamic instability were found to have a small hyperdynamic ventricle on TEE. Of these four patients, three had pulmonary artery occlusion pressure greater than 20 mm Hg. A study by Poelaert et al.²⁰ that evaluated the diagnostic value of TEE compared with pulmonary artery catheterization showed that pulmonary artery catheterization failed to diagnose the presence of hypovolemia in 44% of patients when TEE showed systolic obliteration of the left ventricular cavity, supporting a diagnosis of hypovolemia. TTE and TEE have been shown to play a key role in making the diagnosis of hypovolemia and left ventricular dynamic obstruction, leading to a dramatic impact on therapy.^{19,21,22,82–85}

Assessment of Pulmonary Artery Pressure

Pulmonary hypertension is common in critically ill patients and is a manifestation of various pulmonary, cardiac, and systemic processes. Pulmonary hypertension is said to be present when systolic pulmonary pressure is greater than 35 mm Hg, diastolic pulmonary pressure is greater than 15 mm Hg, and mean pulmonary pressure is greater than 25 mm Hg.⁵⁹ Many echocardiographic methods have been validated for noninvasive estimation of pulmonary artery pressure.^59,86 These methods can be helpful in the ICU. Systolic and diastolic pulmonary artery pressures are determined from the tricuspid and pulmonary regurgitation velocities (some degree of regurgitation is essential to be able to obtain a Doppler signal and subsequently determine pulmonary artery pressure). Tricuspid regurgitation is present in more than 75% of healthy adults⁵⁹ and in approximately 90% of critically ill patients.⁸⁷ Peak tricuspid regurgitation velocity, usually obtained by continuous wave Doppler from the right ventricular inflow or the apical four-chamber view position, reflects the pressure difference during systole between the right ventricle and the right atrium (Figure W2-12).^88–90 Peak systolic pulmonary artery pressure is determined from the peak tricuspid regurgitation Doppler velocity using the modified Bernoulli equation⁹¹: ΔP = 4 × (peak tricuspid regurgitation velocity)². To this peak systolic pressure gradient between right ventricle and right atrium is added the estimated right atrial pressure (see previous section) to obtain the peak right ventricular systolic pressure. In the absence of pulmonic stenosis or right ventricular outflow obstruction, peak right ventricular systolic pressure is equal to systolic pulmonary artery pressure (see Figure W2-12). Echocardiography also can determine diastolic pulmonary artery pressure by applying the modified Bernoulli equation using the regurgitant Doppler velocity of the pulmonary valve to obtain the gradient between the pulmonary artery and the right ventricle at end diastole. To this is added the estimated right atrial pressure (equivalent to right ventricular end-diastolic pressure in the absence of tricuspid stenosis) to obtain end-diastolic pulmonary artery pressure: end-diastolic pulmonary artery pressure = 4 × (peak pulmonary regurgitation velocity)² + estimated right atrial pressure. Approximately 70% of critically ill patients have an adequate Doppler signal of pulmonic insufficiency for this calculation.⁹² Tricuspid and pulmonary regurgitation are present at the same time in more than 85% of subjects.⁹³

Figure W2-12 Calculation of systolic pulmonary artery pressure (PAP). A, Color Doppler transthoracic apical four-chamber view showing a significant tricuspid regurgitation (TR) jet from right ventricle (RV) to right atrium (RA). The peak tricuspid regurgitation velocity is measured by placing the continuous wave Doppler in the center of the tricuspid regurgitation jet (arrow). B, Spectral continuous wave Doppler profile of the tricuspid regurgitation jet. Peak tricuspid regurgitation velocity (3.41 m/s) and peak systolic pulmonary artery pressure gradient (46.5 mm Hg) can be obtained with this modality. C, Peak systolic pulmonary artery pressure also can be determined from the peak tricuspid regurgitation Doppler velocity using the modified Bernoulli equation: ΔP = 4 × (peak tricuspid regurgitation velocity)². To this peak systolic pressure gradient between right ventricle and right atrium is added the estimated right atrial pressure (determined to be 10 in this example) to obtain the peak right ventricular systolic pressure. In the absence of pulmonic stenosis or right ventricular outflow obstruction, peak right ventricular systolic pressure is equal to systolic pulmonary artery pressure. LA, left atrium; LV, left ventricle. (See Color Section in this text.)

Valvular Regurgitation and Prosthetic Valve Dysfunction

In a patient with unexplained hemodynamic instability and a grossly normal TTE examination, performance of subsequent TEE is important to rule out the presence of significant undetected valvular pathology. Common valvular pathologies that can be missed are mitral regurgitation and prosthetic valve dysfunction. In some situations, TTE may provide better imaging than TEE for evaluation of anterior structures such as the aortic valve (native or prosthetic) and for Doppler measurements. TEE is clearly superior to TTE for evaluation of mitral valve pathologies (native and prosthetic). In a study of ICU patients by Alam,¹⁶ TTE compared with TEE was shown either to miss or to underestimate the severity of regurgitation of St. Jude and bioprosthetic valves in the mitral but not in the aortic position.

With acute severe mitral regurgitation, the diagnosis may be clinically difficult because the murmur is often of short duration and low intensity (because of rapid pressure equalization between the left ventricle and the relatively noncompliant left atrium). By TTE, the size of the regurgitant jet in acute mitral regurgitation may appear small and lead to underestimation of severity.⁹⁹ Because of its close anatomic proximity, TEE provides a much more precise evaluation of the degree of mitral regurgitation (Figure W2-13) and provides crucial diagnostic information regarding the cause for mitral regurgitation. The diagnosis of acute mitral regurgitation represents a medical emergency that may necessitate urgent surgery, so the threshold to perform a TEE when this entity is suspected should be low.^8,10,97 Also, several investigators have confirmed the superior accuracy, sensitivity, and reliability of TEE over TTE for dysfunction of mitral prostheses, in which ultrasonic shadowing of the left atrium often occurs with the standard transthoracic studies.^100–103 TEE may be especially useful to detect obstruction of prosthetic valves from thrombus (Video W2-3).

Figure W2-13 Severe mitral regurgitation (MR). Transesophageal five-chamber view shows severe mitral regurgitation with a large regurgitant jet (arrow) going far posteriorly in the left atrium (LA). In this case, systolic flow reversal in the pulmonary veins (another echocardiographic sign of severe mitral regurgitation) also was present (not shown on this picture). Because of its close anatomic proximity, transesophageal echocardiography is an excellent tool for the precise evaluation of the degree of mitral regurgitation. LV, left ventricle; RA, right atrium; RV, right ventricle. (See Color Section in this text.)

Evaluation of the Pericardial Space

Echocardiography is an essential instrument for the diagnosis of pericardial disease. In the ICU, the most common clinical indication for assessment of the pericardial space is suspected tamponade. The pericardium is a potential space that can become filled with fluid, blood, pus, or uncommonly, air. Presence of fluid in this space is detected as an echo-free space. Pericardial fluid usually is detected easily with TTE. The parasternal long-axis and short-axis views and the apical views usually reveal the effusion (Figure W2-14). In many critically ill patients with suboptimal TTE image quality, the subcostal view is often the only adequate window available to detect the presence of a pericardial effusion. In these ICU patients with poor acoustic windows and in the post–cardiac surgical setting, TEE may be needed to assess the pericardial space adequately.

Figure W2-14 Large pericardial effusion. Transthoracic parasternal short-axis view shows a large, predominantly echo-free space around the left ventricle (LV). This space represents fluid in the pericardium. In this case, the large circumferential pericardial effusion was bloody at pericardiocentesis. Particulate matter (e.g., fibrin and clots) can be visualized as denser echoes around the heart and floating in the effusion.

In addition to assisting in the diagnosis of pericardial effusion and tamponade, 2D echocardiography can assist in its drainage, as pericardiocentesis can be performed safely under 2D echocardiographic guidance.^107,108 By determining the depth of the effusion and its distance from the site of puncture, it is possible to optimize the needle placement. Echocardiography also can be used for immediate monitoring of the results of the pericardiocentesis.

Cardiac Tamponade in the Intensive Care Unit

The most common causes of cardiac tamponade in the ICU are listed in Box W2-5. Echocardiographic 2D signs of tamponade are a direct consequence of increased pericardial pressure, leading to diastolic collapse of one or more cardiac chambers (usually on the right side first) (Figure W2-15). Usually, collapse of the right ventricular free wall is seen in early diastole, and right atrial wall collapse is seen in late diastole.¹⁴ This latter sign is sensitive but not specific for tamponade. It is, however, specific for a hemodynamically significant effusion if the right atrial collapse lasts longer than one-third of the R-R interval.^14,109 In the presence of a massive effusion, the heart may have a “swinging” motion in the pericardial cavity. This finding is not always present in cardiac tamponade, because the amount of fluid in the pericardial space may be small but still cause a tamponade physiology, depending on the acuity with which the effusion accumulates and the compliance of the pericardium. In poststernotomy patients, tamponade may be missed by TTE (even in cases in which imaging quality seems adequate) because hematomas causing selective cardiac chamber compression are often in the form of loculated clots located in the far field of the ultrasound beam in the posterior heart region (even when the anterior pericardium is left open).¹¹⁰ The right atrium and right ventricle may be spared in such cases secondary to postoperative adhesions or tethering of the right ventricle to the chest wall anteriorly.¹¹⁰

Figure W2-15 Cardiac tamponade. Transesophageal four-chamber view (A) shows the presence of a large effusion that severely compresses the right atrium (RA) and right ventricle (RV), which appear slitlike. The left ventricle (LV) also is small because of indirect compression and underfilling. Transgastric short-axis view (B) of the same patient shows the large pericardial effusion and severely compressed ventricular chambers. This postcardiotomy patient was in profound shock and was brought back to the operating room emergently for reexploration and drainage of the effusion. LA, left atrium.

Box W2-5

Most Common Causes of Cardiac Tamponade in the Intensive Care Unit

Myocardial or coronary perforation secondary to catheter-based intervention (i.e., after intravenous pacemaker lead insertion, central line placement, or percutaneous coronary interventions)

Compressive hematoma after cardiac surgery

Proximal ascending aortic dissection

Blunt or penetrating chest trauma

Complication of myocardial infarction (e.g., ventricular rupture)

Uremic or infectious pericarditis

Pericardial involvement by metastatic disease or other systemic processes

Another (indirect) sign of a hemodynamically significant pericardial effusion on 2D imaging is plethora of the inferior vena cava with blunted respiratory changes.¹ The latter sign is less valuable in mechanically ventilated patients, because they often have a stiff, dilated inferior vena cava even in the absence of a pericardial effusion (Video W2-4).

Doppler findings of cardiac tamponade are based on characteristic changes in intrathoracic and intracardiac hemodynamics that occur with respiration. Because of the principle of ventricular interaction, mitral inflow velocity (E wave) decreases after inspiration and increases after expiration. Reciprocal changes occur with respect to tricuspid inflow velocity. With tamponade, the exaggerated inspiratory-expiratory variation of the inflow velocity (E wave) over one respiratory cycle should be greater than 40% on the left and greater than 80% on the right.¹¹¹ In critically ill patients, however, mechanical ventilation, bronchospasm, significant pleural effusion, and respiratory distress can alter intrathoracic and intracardiac hemodynamics and make these Doppler findings less reliable. A significant pleural effusion sometimes causes significant respiratory Doppler variations of the inflow velocities that disappear when the effusion is drained.¹¹² The presence of arrhythmia also makes the Doppler findings difficult to interpret. In some circumstances, echocardiographic signs of tamponade may be subtle or absent, so one must keep in mind that the diagnosis of tamponade remains a clinical one and that the echocardiographic signs must be analyzed in conjunction with the clinical findings.

Complications after Cardiac Surgery

Bedside echocardiography has proved to be of particular value in the critical care management of patients with hemodynamic instability after cardiothoracic operations.^{7,8,83,113–115} TTE is often severely limited in this group of patients.^5,8 TEE is the modality of choice in this setting because it provides detailed information that can help determine the cause of refractory hypotension. The most frequent echocardiographic diagnoses encountered in these patients are left ventricular or right ventricular failure, tamponade, hypovolemia, and valvular dysfunction. Schmidlin et al.¹¹⁶ studied 136 patients after cardiac surgery and showed that a new diagnosis was established or an important pathology was excluded in 45% of patients undergoing TEE. A therapeutic impact was found in 73% of cases. The main indications for TEE in this study were control of left ventricular function (34%), unexplained hemodynamic deterioration (29%), suspicion of pericardial tamponade (14%), cardiac ischemia (9%), and “other” (14%). Reichert et al.¹¹³ performed TEE in hypotensive patients after cardiac surgery. Left ventricular failure was found in 27% of patients, hypovolemia in 23%, right ventricular failure in 18%, biventricular failure in 13%, and tamponade in 10%. Comparison with hemodynamic parameters showed agreement on diagnosis (hypovolemia versus tamponade versus cardiac failure) in only 50% of the cases. Echocardiography identified two cases of tamponade and six of hypovolemia that were not suspected based on standard hemodynamic data. In five patients with hemodynamic findings suggesting tamponade, unnecessary reoperation was prevented because TEE ruled out this diagnosis. Costachescu et al.²² also showed the superiority of TEE compared with conventional monitoring with pulmonary artery catheterization in diagnosing and excluding significant causes of hemodynamic instability in postoperative cardiac surgical patients.

Descriptions of the echocardiographic findings of left ventricular dysfunction, tamponade, hypovolemia, and valvular dysfunction were described earlier in this chapter.

Infective Endocarditis

Occurrence of infective endocarditis in patients hospitalized in an ICU is common. It is often in the differential diagnosis of febrile patients in the ICU. Infective endocarditis was the second most common indication for performance of an echocardiogram among centers reporting their experience, as summarized in a review article by Heidenreich.⁵ Nearly all critically ill patients are at risk for iatrogenic infection, bacteremia, and subsequent endocarditis because of the presence of multiple indwelling catheters, severe underlying diseases, malnutrition, and prolonged mechanical ventilation. Classic clinical findings suggesting endocarditis¹⁰⁶ are uncommon in this patient population. Echocardiography is the test of choice for the noninvasive diagnosis of endocarditis. Fowler et al.¹¹⁷ studied patients with Staphylococcus aureus bacteremia referred for TEE and showed that endocarditis ultimately was diagnosed in 25%. Only 7% of these patients had physical findings suggesting endocarditis before TEE. Absence of clinical stigmata is especially likely if the infection presents acutely. Because the consequences of untreated endocarditis are devastating and often ultimately fatal, it is important that the infection and its complications be recognized promptly and treated appropriately.⁵⁹

The echocardiographic features typical for infective endocarditis are^59,118 (1) an oscillating intracardiac mass on a valve or supporting structure or in the path of a regurgitant jet or an iatrogenic device, (2) abscesses, (3) new partial dehiscence of a prosthetic valve, or (4) new valvular regurgitation. Sensitivity for the echographic diagnosis of endocarditis is 58% to 62% for TTE and 88% to 98% for TEE.^119,120 TEE is particularly useful for detecting small vegetations¹²¹ and detecting vegetations on prosthetic valves. TEE also has been shown to be superior to TTE for diagnosing complications of endocarditis such as aortic root abscess, fistulas, and ruptured chordae tendineae of the mitral valve.¹⁶ Among ICU patients, sensitivity of TTE for the diagnosis of endocarditis is often poor because the quality of the transthoracic study is commonly suboptimal. The sensitivity of TEE for suspected infective endocarditis usually is excellent in the ICU (Figure W2-16). In a study by Font et al.,⁹⁷ a search for vegetations was the indication for 51 (46%) of 112 TEE studies performed for critically ill patients. TEE increased the detection rate by 27% compared with TTE. Suspicion of endocarditis represented 29% of the indications for TEE in a study of ICU patients by Chenzbraun et al.⁸⁵; 9 (27%) of 31 patients with suspected infective endocarditis had a positive study for endocarditis. All positive studies were in patients who had an increased likelihood for infective endocarditis before the examination, as indicated by the presence of fever, positive blood cultures, new-onset murmur, prosthetic valve, or new-onset heart failure (alone or in combination). None of the patients with native valves and no clinical features of endocarditis had a TEE study diagnostic of infective endocarditis, and in none of them was the diagnosis of infective endocarditis made later. The findings from this study indicate that TEE is not useful as a screening procedure for infective endocarditis in septic patients without high clinical likelihood for endocarditis. The clinical probability of endocarditis should guide the use of TEE. Clinical risk factors considered high risk included intracardiac prosthetic material, positive blood cultures (in particular S. aureus), evidence of peripheral emboli, and history of previous endocarditis^8,16 (Video W2-5).

Figure W2-16 Infective endocarditis of the aortic valve. A 55-year-old patient was admitted to the ICU with fever, chills, hypotension, and respiratory distress for which he had to be intubated. He had 4/4 positive blood cultures for Staphylococcus aureus. Transthoracic echocardiography was performed initially, but the quality was suboptimal, and no definite conclusion could be reached. Subsequent transesophageal echocardiography revealed a large vegetation on the left aortic coronary cusp as seen in the midesophageal view at 120 degrees (A). Color Doppler examination (B) revealed the presence of associated severe aortic regurgitation (AR). The patient was treated with antibiotics and emergent aortic valvular surgery. LA, left atrium; LV, left ventricle. (See Color Section in this text.)

As concluded by Colreavy et al.,⁸ performance of TEE in the ICU for suspicion of infective endocarditis should be (1) for cases associated with a clinical likelihood of endocarditis and a negative TTE examination, (2) for suspected prosthetic valve endocarditis, (3) for assessment of complications in known cases of endocarditis, and (4) for cases of S. aureus bacteremia when the source is unknown or blood cultures remain positive despite antibiotic therapy. When assessing a patient for infective endocarditis by echocardiography, one must keep in mind the noninfectious causes of vegetations that may result from tumors, myxomatous degeneration, marantic endocarditis, Lambl’s excrescences, valve thrombus, and suture material in patients with repaired native or prosthetic valves.

Assessment of the Aorta

In the ICU, use of bedside echocardiography for assessment of suspected aortic pathologies provides many advantages over CT or aortography: There is no need for IV contrast administration, there may be less time delay, there is no need for transportation of a critically ill patient, and cardiac morphology and function can be evaluated at the same time.³ For many years, aortography has been the gold standard for the investigation of suspected injuries of the aorta.³ The advent of noninvasive modalities such as CT, magnetic resonance imaging (MRI), and TEE with their excellent sensitivity and specificity to diagnose aortic pathologies has decreased the need for aortograms.

Suspected aortic pathologies can be encountered in different ICU settings. The aorta may have to be imaged to rule out dissection, rupture, aneurysm, aortic debris, or aortic abscess. TTE is a good initial imaging modality for evaluation of the proximal aorta (ascending aorta and arch),⁵⁹ but the descending thoracic aorta cannot be adequately assessed and visualized with this modality. Because of the close anatomic relationship between the thoracic aorta and the esophagus, TEE allows optimal visualization of the entire thoracic aorta (Figure W2-17). As described earlier, there exists a blind spot in the distal portion of the ascending aorta and the proximal portion of the transverse aorta where imaging can be suboptimal.^122,123

Figure W2-17 Dissecting thoracic aortic aneurysm. A 65-year-old patient presented to the emergency department with severe ripping chest pain radiating to the back. The initial electrocardiogram was unremarkable, and the chest x-ray showed a widened mediastinum. The patient underwent transesophageal echocardiography, which revealed the presence of a large dissecting aneurysm of the descending thoracic aorta. The short-axis view (A) revealed the presence of a large aneurysm with a true and a false lumen. The false lumen was filled with thrombus (arrow). On the longitudinal view with color Doppler (B), blood flow in the true lumen is visualized. The patient was taken emergently to the operating room. (See Color Section in this text.)

Assessment for Intracardiac and Intrapulmonary Shunts

In critically ill patients, clinical suspicion for an intracardiac or intrapulmonary shunt most often is raised in the context of unexplained embolic stroke or refractory hypoxemia. In such cases, the presence of a right-to-left shunt must be excluded. Common origins of right-to-left shunt are atrial septal defect or patent foramen ovale at the cardiac level⁵ and arteriovenous fistula at the pulmonary level.⁵ To be able to detect the presence of such a shunt at the bedside, a contrast study often is needed, because the shunt is usually not well visualized with 2D echocardiography alone. Color-flow imaging increases the detection rate of intracardiac shunt to some extent, but usually only when the shunt is large. Accordingly, a contrast study should be performed routinely as part of a TEE or TTE examination when evaluating a patient with unexplained embolic stroke or refractory hypoxemia in the ICU. For this purpose, agitated saline contrast is usually used. Approximately 0.5 mL of air is mixed with 10 mL of normal saline and is vigorously agitated back and forth between two syringes connected to the patient by a three-way stopcock. After an adequate echocardiographic view of the right and left atrial cavities has been obtained, the agitated saline is forcefully injected IV. After injection, the contrast is seen in the vena cava, right atrium, right ventricle, and pulmonary artery. In the absence of a shunt, only a minimal amount of contrast should be seen in the left-sided cavities, because most of the microbubbles from the agitated saline are unable to pass through the pulmonary capillaries. If an intracardiac shunt is present, such as an atrial septal defect or patent foramen ovale, left-sided contrast is observed immediately after right-sided opacification, and the contrast is seen going through the interatrial septum (Figure W2-18). Performance of a Valsalva maneuver by the patient during contrast injection increases the sensitivity of the bubble study to detect right-to-left shunting. In mechanically ventilated patients, a maneuver equivalent to a Valsalva may be performed by inducing sudden release of sustained airway pressure previously achieved by inflating the lungs manually. This maneuver reverses the atrial transseptal gradient and may help uncover a patent foramen ovale that would not have been seen otherwise. Right-to-left shunting also can be caused by the presence of pulmonary arteriovenous fistulas. These often are associated with end-stage liver disease (hepatopulmonary syndrome). With this type of shunt, contrast is seen to appear in the left atrium from the pulmonary veins instead of through the atrial septum; this finding is best detected by TEE, which usually permits visualization of all four pulmonary veins. The characteristic of intrapulmonary versus intracardiac shunt is that there is a longer delay (three to five cardiac cycles) between the appearance of contrast from the right-sided to left-sided cavities in the presence of an intrapulmonary shunt.⁵ Agitated saline is a simple and easy way to use contrast at the bedside.

Figure W2-18 Positive bubble study shows the presence of a right-to-left shunt via a patent foramen ovale (PFO). Transesophageal echocardiography was performed in a patient hospitalized in the ICU for pneumonia. He presented with refractory hypoxemia that was out of proportion to the underlying minor pulmonary process. Transesophageal echocardiography was obtained (multiplane transducer at 98 degrees) and showed the presence of a patent foramen ovale with a significant right-to-left shunt due to elevated right atrial pressure. Soon after contrast injection (A), the bubbles are seen arriving in the right atrium (RA) from the superior vena cava (SVC). A few seconds later (B), a complete opacification of the right atrium is reached, and the bubble contrast is clearly seen shunting through the patent foramen ovale from the right atrium to left atrium (LA).

Other types of intracardiac shunts also can be encountered in the ICU. After myocardial infarction, patients can develop cardiogenic shock due to acute development of a ventricular septal defect and resultant left-to-right shunt. Physical examination and invasive hemodynamic monitoring (pulmonary artery catheterization) sometimes can miss this diagnosis. Echocardiography reveals a disrupted ventricular septum with a high-velocity left-to-right shunt. This kind of shunt usually is well visualized without use of contrast. The diagnosis can be established by 2D and Doppler TTE in approximately 90% of cases.¹³⁴ Penetrating cardiac trauma is often associated with intracardiac and extracardiac shunts, and TEE is becoming the obvious tool for perioperative early identification of occult shunts.¹⁴ Identification of these shunts is paramount in these critically ill patients, because missing them may lead to cardiac tamponade and rapid death. TEE has been shown to be superior to angiography and TTE to visualize these lesions.^135–137

Source of Embolus

In the setting of acute unexplained stroke, echocardiography often is required to determine whether a potential embolic source of cardiac origin is present. TEE is the modality of choice for this purpose. Possible cardiac sources of emboli to the arterial circulation include left atrial or appendicular thrombus, left ventricular thrombus, thoracic atheromatosis, and right-sided clots (right atrium, right ventricle, vena cava) combined with a right-to-left intracardiac shunt (leading to a paradoxical embolus). Cardiac tumors and vegetations are other potential sources of emboli from cardiac origin that must be considered.

When cardioversion is considered for a critically ill patient with atrial fibrillation or flutter, performance of TEE is helpful in evaluating the left atrium and appendage for the presence of thrombus (Figure W2-19). If no intracardiac clots are documented, cardioversion can be performed with minimal embolic risks.

Figure W2-19 Large clot in the left atrial wall and left atrial appendage (LAA). A 72-year-old patient hospitalized in the ICU for urosepsis developed rapid atrial fibrillation. She was initially anticoagulated and rate-controlled. Despite resolution of the septic picture, she remained in atrial fibrillation 4 days after its onset. The patient underwent transesophageal echocardiography before undergoing a planned electrical cardioversion. The midesophageal view with multiplane transducer at 70 degrees revealed the presence of a large clot (3.35 × 1.45 cm) in the posterolateral wall of the left atrium (LA) extending into the LAA. With these findings, the patient’s anticoagulation regimen was intensified, and the cardioversion was not performed. LV, left ventricle.

Central Line Placement

Central venous catheterization is performed frequently in critically ill patients. Placement of a central venous catheter is not without risk and can be associated with adverse events that are hazardous to patients and expensive to treat.^162–164 Complications can be seen in 15% to 20% of cases.^165–167 As described in a review by McGee and Gould,¹⁶⁸ complications related to central venous line placement are most often mechanical (arterial puncture, local hematoma, hemothorax, pneumothorax), infectious (catheter colonization and related bloodstream infection), and thrombotic. Complications are influenced by patient factors (obesity, coagulopathy, previous failed catheterization), site of attempted access, and operator experience.¹⁶⁹ As previously reported, only approximately 38% to 65% of patients are cannulated on the first attempt using a blind method.^170,171

The use of ultrasound guidance during central venous catheterization has been well shown to reduce the risk of complications, mostly so for the internal jugular route. Ultrasound guidance also speeds catheter placement, decreases the number of attempts before successful placement, and improves the overall rate of successful placement. Ultrasound can be used to help localize and define the anatomy of the vein, with subsequent placement of the central venous catheter by the standard use of anatomic landmarks at the site identified by ultrasound, with the knowledge that a vein is present, patent, and of adequate size. Ultrasound also can be used to provide real-time 2D ultrasound guidance to locate the vein and subsequently introduce the needle through the skin and into the vessel. Multiple studies have reported the superiority of ultrasound-assisted cannulation of the internal jugular vein in ICU patients, compared with the external landmark–guided technique.^170–172 Trials looking at ultrasound guidance after failure by the landmark method reported success rates ranging from 33% to 100%.^{169,173–175} A meta-analysis¹⁷⁵ of the literature comparing guidance using anatomic landmarks only versus guidance using ultrasound for the placement of central venous catheters indicates that ultrasound guidance significantly decreases placement failure by 64%, decreases related complications by 78%, and decreases the need for multiple placement attempts by 40%. Data showing superiority of the ultrasound guidance technique are consistent and strong for the internal jugular vein approach but less so for subclavian venous catheterization.^175–177

Some patients can be identified in whom cannulation may be more difficult or in whom consequences of a complication could be more serious.¹⁶⁹ In these patients (Box W2-6), central venous cannulation may be laborious and risky, and ultrasound guidance should be considered. Hatfield and Bodenham¹⁶⁹ showed the benefit of portable ultrasound when central venous access was difficult. As suggested by this study and others,¹⁷⁸ ultrasound guidance is particularly beneficial when used in difficult cases or when a competent operator fails after a few attempts using surface landmarks.

Ultrasound guidance is useful for operators with varying levels of experience.^169,175 The technique is easy to learn and can be self-taught with some practical assistance from radiologists or other experienced sonographers.^169,179,180 Familiarity with the anatomy and equipment is easy to obtain safely at the bedside.

Most large vessels that are catheterized usually can be imaged by ultrasound. Different types of ultrasound modalities can be used to help guide central vessel cannulation, including 2D ultrasound, Doppler transducer, Doppler with the probe in the needle, and fingertip pulse Doppler. With 2D imaging, fluid such as blood in vessels is black because there is nearly complete transmission of ultrasound.¹⁶⁹ Color Doppler mode helps delineate the flow patterns in vessels. Doppler-only equipment that provides no images has shown equivocal results in studies of vascular access.^176,181

With 2D imaging, arteries are characteristically small, pulsatile, and difficult to compress with the probe.¹⁶⁹ Veins are usually larger, are nonpulsatile (except in the presence of severe tricuspid regurgitation), are easily compressible, and distend when the patient is placed with the head down or when a Valsalva maneuver is performed.¹⁶⁹

Vessels can be examined in the transverse and longitudinal views. The transverse view permits identification of the vein and arteries based on the sonographic characteristics mentioned earlier and clarifies their positions relative to one another (Figure W2-21). The transverse and longitudinal views enable the sonographer to monitor in real time the passage of the needle through the skin and the anterior vessel wall. Ultrasound guidance also ensures detailed and accurate control of the needle (Figure W2-22).¹⁶⁹

Figure W2-21 Transverse view of normal anatomy of the right internal jugular (RIJ) vein and right common carotid artery (RCCA). Ultrasound examination helps determine the anatomic relationship, size, and patency of the vessels. Knowledge of these important vessel characteristics helps determine if the anatomy is suitable for central vein catheterization at a low risk. If the vessel anatomy is normal and the operator is experienced, subsequent venous catheterization can be done by the surface landmark technique or under real-time ultrasound guidance. If high-risk characteristics are identified (see Box W2-6), however, real-time ultrasound guidance (or selection of a different access site) would be preferred.

(Courtesy Dr. Kurian Puthenpurayil.)

Figure W2-22 Transverse view of left carotid artery and left internal jugular (IJ) vein. The distance from the skin to the anterior wall of the vein is measured before insertion of a central venous catheter (A). Knowledge of this distance prevents the operator from going too deep with the needle when searching for the vein; this helps decrease the incidence of pneumothorax. After insertion, the catheter position in the jugular vein is confirmed (B).

During vessel examination, the sonographer specifically should assess the presence and patency of the vein (Figure W2-23), the distensibility and compressibility of the vein, the position of the vein relative to the surrounding arteries (Figure W2-24), and the presence of a thrombus in the vein (Figure W2-25).¹⁶⁹ Ultrasound identification of certain anatomic characteristics such as small vessel size (<5 mm), intraluminal thrombus, and anterior location of the artery relative to the vein helps the physician identify unfavorable vessel anatomy and choose another catheterization site. A study by Levin et al.¹⁸² showed that 2D ultrasound guidance for the insertion of radial artery catheters was easy to use and increased the rate of success of insertion at first attempt. It was determined to be a useful adjunct to arterial catheter insertion. More studies are needed in the use of ultrasound for cannulation of peripheral arterial conduits.

Figure W2-23 Transverse view of the left carotid and internal jugular (IJ) vein in a spontaneously breathing patient sitting in bed at a 30-degree angle. The patient was febrile and dehydrated. A near-total collapse of the vein (which is of small caliber) can be appreciated on inspiration (B) compared with expiration (A).

Figure W2-24 Transverse view of left internal carotid artery and left internal jugular (IJ) vein. Notice the relative position of the jugular vein directly overlying the carotid artery. This type of anatomy is common on the left side and, when present, significantly increases the risk of procedure failure or arterial puncture.

Figure W2-25 Transverse (A) and longitudinal (B) views of the right carotid artery and right internal jugular vein. Complete thrombosis of the right internal jugular (IJ) vein can be appreciated. Notice the small caliber of the thrombosed vein and the increased echogenicity of the thrombotic material within it. The vessel could not be compressed by probe pressure.

Assessment of Pleural Effusions and Intraabdominal Fluid Collections

In critically ill patients, atelectasis and pleural effusions are frequent and often are present at the same time. Patients in the ICU are most often supine, and chest x-rays performed in this position offer limited sensitivity for the diagnosis of pleural effusion.¹⁸³ In many instances, neither atelectasis nor infiltration can be differentiated from pleural effusion. An alternative diagnostic method is needed to provide better results. Decubitus chest radiographs may show if fluid is free flowing, but this approach cannot localize or characterize the effusion precisely. CT of the chest shows the amount and distribution of fluid and is superior to plain lateral decubitus films. CT also can differentiate fluid from atelectasis and reveal information about the lung parenchyma. Chest CT requires transport to the radiology suite, however, which can be hazardous in unstable critically ill patients. Ultrasound examination of the pleural space has proved to be valuable for diagnosis of effusion.^184–188 The value of ultrasound for localizing fluid before catheter drainage or simple thoracentesis is well recognized. Ultrasound is especially valuable for localizing loculated or small effusions before a drainage procedure. In mechanically ventilated patients, blind thoracentesis can be hazardous, especially if the effusion is small or if the patient is on a high level of positive end-expiratory pressure.¹⁸⁹ Lichtenstein et al.¹⁸⁹ evaluated the feasibility and safety of ultrasound-aided thoracentesis in 40 mechanically ventilated patients. No complications occurred in the 45 ultrasound-aided thoracenteses, all performed by ICU physicians.

Basic skill required to detect a pleural effusion may be acquired in minutes and improves with experience.¹⁹⁰ In most instances, the pleural tap does not have to be done under real-time ultrasound guidance. A critically ill patient first must be positioned adequately on the back or on the side. Scanning of the pleural space is performed with the ultrasound probe. The probe must be oriented upward and downward, laterally and medially, and anteriorly and posteriorly so as to obtain a complete anatomic assessment of the area. The pleural fluid is usually hypoechogenic and appears black. The surrounding solid structures (soft tissue, diaphragm) and organs (lung, liver, heart, spleen) are visualized as structures with different degrees of echogenicity around the effusion (Figure W2-26). The presence of aerated lung causes airy artifacts. Ribs usually yield artifactual anechoic images. When the effusion has been well assessed, one must determine the feasibility of safely doing a thoracentesis. One must check for the absence of interposition of lung, heart, liver, or spleen during the respiratory cycle¹⁸⁹ to avoid puncturing these organs, which potentially can cause catastrophic complications. When an optimal and safe position for thoracentesis has been determined, the skin should be marked and disinfected, and the patient should remain in the exact same position as was used during the ultrasound examination. Optimally, the puncture should be done within seconds to minutes of the marking.

Figure W2-26 Transverse view of a right pleural effusion. Collapsed atelectatic lung is well visualized “floating” in the effusion.

(Courtesy Dr. Kurian Puthenpurayil.)

The same diagnostic and therapeutic procedures described earlier can be applied for intraabdominal fluid collections in a critically ill patient. Evaluation for intraabdominal fluid collection or abscess is restricted to areas that are not impeded by gas-filled structures¹⁹¹ and include the regions around the liver and gallbladder, spleen, kidneys and lateral retroperitoneal areas, and pelvis around the uterus and bladder.¹⁹¹ Fluid that does not change shape with probe pressure or patient positioning most likely represents a loculated collection.¹⁹¹ Echogenic material and diffuse echoes on ultrasound within a fluid collection suggest the presence of particulate matter (e.g., fibrin or clots) and may represent an exudate or blood collection. As with pleural effusions, intraabdominal fluid collections can be percutaneously sampled or drained safely at the bedside under real-time ultrasound guidance (Figure W2-27).

Figure W2-27 Transverse view of a left lower quadrant abdominal collection. Echogenic, particulate material can be seen floating in the collection. Loops of bowel also are well visualized. A 6F catheter was inserted under ultrasound guidance to drain the collection, which was found to be chylous. Fluid collection with echogenic material and diffuse echoes on ultrasound are suggestive of particulate matter (e.g., fibrin or clots) and may represent an exudate or blood collection.

(Courtesy Dr. Kurian Puthenpurayil.)

Urinary Bladder Scan

Bladder scanning devices are portable units that can provide a measurement of urine volume in the bladder (Figure W2-28) and avoid bladder overdistention and reduce the need for unnecessary catheterization.^191,192 Studies have shown that frequent catheterization is a major risk factor for urinary tract infections that can be costly to medical centers.^193–195 Use of a portable bladder scanning device to reduce the incidence of nosocomial urinary tract infections was described by Moore and Edwards.¹⁹⁶ Bedside ultrasound assessment of volume in the urinary bladder also can be helpful to evaluate oliguria or anuria to rule out obstruction of the urinary catheter.

Figure W2-28 Urinary bladder. Suprapubic transverse view of a full urinary bladder (A). This “square” appearance of the bladder with a concave superior wall is typical of a moderately full bladder. When overdistended (i.e., in the presence of a low urinary tract obstruction), the bladder is large and adopts a round, globular shape (not shown). Suprapubic transverse view of an empty bladder (B). When empty, the bladder can become small and commonly may be difficult to identify.

(Courtesy Dr. Kurian Puthenpurayil.)

Focused Assessment of the Trauma Patient

Since the early 1990s, bedside ultrasound has been used in the United States as an additional diagnostic modality for use in determining the presence of intraabdominal injury after blunt trauma.¹⁹⁷ It is performed in the trauma bay during the secondary survey (as described in Advanced Trauma Life Support) or as part of the primary survey in hemodynamically unstable patients.^{191,198–202} The focused assessment for sonographic examination of trauma (FAST) should be done with a specific purpose, usually identification of hemoperitoneum, hemothorax, or tamponade.¹⁹¹ FAST seeks to determine the presence of fluid in four areas: (1) the subxiphoid region in the pericardial sac, (2) the right upper quadrant in Morison’s pouch, (3) the left upper quadrant in the splenorenal recess, and (4) the pelvis in the pouch of Douglas or rectovesical space (Figure W2-29).¹⁹ Because the FAST examination is noninvasive and quickly performed at the bedside, it is ideal for detecting intraabdominal injury in the resuscitation area. It has now been incorporated into the trauma resuscitation algorithm of most level I trauma centers in the United States.^191,203

Figure W2-29 Focused assessment for sonographic examination of trauma (FAST). The FAST examination seeks to determine the presence of fluid in four areas: (1) the subxiphoid region in the pericardial sac, (2) the right upper quadrant in Morison’s pouch, (3) the left upper quadrant in the splenorenal recess, and (4) the pelvis in the pouch of Douglas or rectovesical space.

Use of the FAST examination has been shown to diminish the need for more invasive diagnostic measures such as diagnostic peritoneal lavage and subsequent exploratory laparotomy.^204,205 The FAST examination has been shown to be most accurate when performed for evaluation of hemodynamically unstable patients.^{203,206–208} Studies have suggested that its use as a screening tool for blunt abdominal injury in hemodynamically stable trauma patients may result in under diagnosis of intraabdominal injuries.^202,203,209

Intraaortic Balloon Counterpulsation

Bedside TEE may be helpful in different aspects of intraaortic balloon counterpulsation management. Before insertion, TEE can rule out the presence of significant aortic regurgitation, which would represent a contraindication to intraaortic balloon counterpulsation use. After insertion, TEE can confirm the position of the intraaortic catheter in the descending thoracic aorta, ensure correct functioning of the balloon (visualization of inflation and deflation), and rule out the presence of important complications of aortic catheter insertion (e.g., aortic dissection). TEE also may be used for monitoring of the ventricular function while separating the patient from the intraaortic balloon counterpulsation device.

Ventricular Assist Devices

Different complications are likely to occur after ventricular assist device implantation, such as bleeding and hemodynamic instability. Maintenance of ventricular assist device flow is a key indicator of the overall status of the system. In the postoperative period, low ventricular assist device flow is usually due to hypovolemia and right ventricular dysfunction. TEE can be helpful for the diagnosis and monitoring of both of these conditions. Right ventricular failure has been shown to occur in approximately 20% to 25% of patients being supported with an isolated left ventricular assist device.²¹⁰ With prosthetic circulatory support devices, there can be dramatic changes in ventricular volumes and hemodynamic conditions and substantial direct and indirect changes to the contralateral ventricle due to ventricular interactions. TEE can help the clinician monitor and understand these ventricular interactions.²¹¹ It also can help assess adequacy of flow and the patency of the inflow and outflow cannulas to eliminate the presence of a thrombus and collapse or displacement of the cannulas. It also can motivate an urgent return to the operating room if a cardiac tamponade is diagnosed. If hypoxemia supervenes in the ICU, the presence of a patent foramen ovale has to be ruled out. For patients placed on extracorporeal membranous oxygenation support, bedside TEE also can be used to monitor ventricular function during weaning of the circulatory assistance.

Performance of Bedside Ultrasonography by the Intensivist

In acute situations in the ICU, it may be difficult to have a cardiologist or sonographer available on immediate call on a 24-hour basis to perform a bedside ultrasound examination. The value of immediate bedside echocardiography for aiding in diagnosis and management of acute hemodynamic disturbances has been well shown in the literature in the ICU and the emergency department.^212,213

Ultrasound technologies are not exclusive to the radiologist or cardiologist. Appropriately trained emergency department physicians, surgeons, anesthesiologists, and ICU specialists have been using ultrasound devices with great success. Anesthesiologists were instrumental in many of the pioneering studies of TEE in the operating room and ICU.^4,22,214,215 Successful performance of bedside echocardiography by noncardiologist intensivists also has been well shown in the literature.^8,21,216 A study by Benjamin et al.²¹ showed that a limited TEE examination performed and interpreted by intensivists (after training under the supervision of two cardiologists) is feasible and provides rapid, accurate diagnostic information that can have a dramatic impact on the treatment of critically ill patients.²¹ The safety and utility of performance of bedside ultrasound by the intensivist for various other purposes in the ICU (central venous cannulation, thoracentesis, paracentesis) also have been well shown.^170–172189

With the increasing popularity of ultrasound devices—particularly lightweight, portable, hand-held devices—there is controversy regarding the advisability and use of noncomprehensive “goal-directed” examinations performed by clinicians without cardiology or radiology training.¹⁰⁴ Studies with these portable devices that provide basic 2D and Doppler flow imaging showed they can provide important anatomic information^216–220 but that even in highly skilled hands, they may provide suboptimal imaging or diagnostic capabilities in the ICU.²¹⁸ Inappropriate interpretation or application of data gained by a poorly skilled user may result in adverse medical, ethical, and social consequences.¹⁰⁴ To avoid misusing the technology, adequate training is essential.

The era of a technology-extended physical examination²¹⁹ seems to have arrived, and there seems to be a role for a user-specific, focused ultrasound examination.^104,221 An examination said to be “targeted,” “focused,” and “limited” may often equate with “incomplete,” “inadequate,” or “inaccurate.” Training must be individualized and tailored to specific needs, and appropriate user-specific application depends directly on the training and expertise of the user.¹⁰⁴ Provided that adequate expert backup is available, the training of intensivists in performing focused or more comprehensive bedside ultrasound examinations is not only feasible but also can be done safely and rapidly and yield information pertinent to the management of critically ill patients. General guidelines in training for TTE and TEE have been developed by the American Society of Echocardiography in association with the American Heart Association and the American College of Cardiology.²²² Since 1996, the American Society of Anesthesiologists and Society of Cardiovascular Anesthesiologists also have developed practice guidelines for perioperative TEE.²²³ The importance of adequate training and subsequent maintenance of competence cannot be overemphasized; inappropriate use or misapplication potentially could temper the acceptance and limit the value of performance of bedside ultrasonography by the intensivist.

Training of intensivists and emergency department physicians in performance of emergency bedside ultrasonography should provide rapid answers to clinical questions that may strongly affect medical and surgical management decisions. As has been mentioned by different authors,^190,224 training in echocardiography and general ultrasonography should be incorporated into the critical care fellowship, with special emphasis on TEE as part of the training program. It is hoped that critical care and echocardiographic societies will credential such additional training in the near future.

Key Points