Echocardiography in - Europe PMC

Update on Cardiovascular Imaging

Utility of Transesophageal Echocardiography in Infective Endocarditis A Review

Carlos Jessurun, MD Andres Mesa, MD Susan Wilansky, MD

Key words: Echocardiography, transesophageal; echocardiography, transthoracic; endocarditis, bacterial; heart valve diseases; heart valve prosthesis

From: The Department of Adult Cardiology, Texas Heart Institute and St. Luke's Episcopal Hospital, Houston, Texas 77030 Section editors: Warren H. Moore, MD Barry D. Toombs, MD Susan Wilansky, MD Address for reprints: Susan Wilansky, MD, 6624 Fannin, Suite 2480, Houston, TX 77030

98

Despite recent diagnostic and therapeutic advances, infective endocarditis continues to be a very serious illness, with high patient morbidity and mortality rates. The diagnosis of infective endocarditis has been based primarily on clinical signs and positive blood cultures. Echocardiography is currently recognized as the technique of choice for the detection of valvular vegetations, which are the hallmark of endocarditis. We briefly review the use of echocardiography in the diagnosis of suspected infective endocarditis, with emphasis on transesophageal echocardiography. High-resolution imaging of the cardiac valves with transesophageal echocardiography has proved to be invaluable in the management of infective endocarditis. (Tex Heart Inst J 1996;23:98107)

I nfective endocarditis continues to be a life-threatening illness with high morbidity and mortality rates. Early diagnosis is crucial but difficult to establish on the basis of clinical observations alone. Transesophageal echocardiography has made possible high-resolution imaging of the cardiac valves, adding a new dimension to the diagnostic approach to infective endocarditis. The infective process begins with the formation of a small, noninfected thrombus on an abnormal endocardial surface. The fibrin matrix of the thrombus becomes secondarily infected by blood-borne agents prone to proliferation in this environment. The enlarging infected thrombus (vegetation) causes disease by multiple mechanisms. The intracardiac, destructive process frequently erodes valve leaflets or perivalvular structures, leading to valve regurgitation and, rarely. stenosis. Abscess and fistula formation and the invasion of adjacent conduction tissue may occur. Sustained, low-grade bacteremia can cause metastatic infection and a deleterious immunologic response. Septic emboli may also result in severe organ damage.' Certain strains of organisms are capable of infecting normal cardiac tissue de novo. Most infections occur on preexisting abnormal structures. including prosthetic valves and native cardiac anomalies. This results in high- to low-pressure pathways such as those occurring in valvular regurgitation and ventricular septal defects.23 The technique of M-mode echocardiography, introduced in the early 1970s by Dillon's group, 1st made the premortem or preoperative diagnosis of vegetations possible. 5 Since then, echocardiography has been increasingly ulsecl in the evaluation of patients with suspected infective endocarditis.&25 With M-mode echocardiography, the detection rate of vegetations has varied from 140S, to 750 o.26 In 1977, Gilbert and coworkers' were among the 1st to describe the 2-dimensional echocardiographic findings of vegetations on precordial echocardiography. The superior spatial resolution of 2-dimensional echocardiography yielded higher detection rates of vegetations compared with the results of M-mode echocardiography.9""6 182( Estimates of sensitivity varied between 40% and 700o.` 10"2-30 The recent development of transesophageal echocardiography (TEE) has ftirther enhanced the capability of visualizing vegetations and abscesses.2 314 In 1976, Frazin and co-authors48 described their initial experience with TEE by use of a single-crystal ultrasound transducer attached to a coaxial cable that was passed into the esophagus. A major breakthrouLgh in TEE came in the early 1980s with the introduction of smaller, phased-array transcducers connected to more flex-

Transesophageal Echocardiography in Infective Endocarditis

l'ohime 21-)', Number 2. 1996

ible endoscopes.4952 Initially, TEE was undertaken with monoplane probes that allowed scanning only in the transverse imaging plane. Later, biplane probes were developed. Currently, biplane probes with spectral and color Doppler are the standard form of technology. Multiplane TEE probes have also become available, allowing stepwise study of a particular area of interest by fine mechanical or electronic rotation of the scanning plane at 180'.559

Diagnostic Criteria The 1st recorded description of a vegetation was made by Riviere in 1646: "In the left ventricle, round corpusculae were found, the larger of which resembled a cloisture of hazelnuts and filled up the opening of the aorta."60 Today, the presence of a vegetation is the most prominent echocardiographic characteristic of infective endocarditis. On echocardiography, a vegetation is a nonuniform mass, usually shaggy or irregular in appearance. In cases of infective endocarditis, the vegetation is either contiguous with a valve leaflet (or prosthetic valve) or on the endocardial surface.273235 Valve ring abscess, in the setting of valvular infection, appears as an abnormal echolucent area within the perivalvular tissue.45 Before echocardiography was developed, the diagnosis of infective endocarditis was based on stringent clinical criteria6' that are still useful in the initial evaluation of patients with suspected infective endocarditis. Bacteremia without an extracardiac source of infection is an important feature, particularly in the presence of known or suspected underlying cardiac structural abnormalities such as a prosthet-

ic valve or congenital heart defect. Lukes and col-

leagues62-4 have pointed out the difficulty in making a firm diagnosis of infective endocarditis because of the limited sensitivity and specificity of the current clinical criteria. Therefore, echocardiographic documentation of vegetations or abscesses should be included as a major diagnostic criterion. Guidelines for the diagnosis of infective endocarditis were published by Lukes and associates63 in 1992 (International Science Conference on Antimicrobial Agents and Chemotherapy). New criteria have recently been proposed,26'M emphasizing the importance of echocardiography in the diagnostic process (Table I). In this approach, echocardiographic findings are as important as blood cultures and clinical findings in the diagnosis of infective endocarditis. Patients who are suspected of having infective endocarditis by clinical criteria should 1st be screened by use of transthoracic echocardiography (TTE) and then, depending on the findings, by TEE (Fig. 1). An abnormal result on a transthoracic echocardiogram is relatively specific for endocarditis, with estimates in the range of 80% to 90%. 10,14,27,29.30,65 In patients exhibiting clinical evidence of infective endocarditis, the overall detection rate for vegetations by lE; is 50%.26 Two-dimensional TFE may yield inadequate images in up to 20% of cases because of chest wall abnormalities, underlying pulmonary disease, obesity, recent thoracic surgery, or the use of mechanical ventilation.26 Transthoracic echocardiography is preferred over TEE for the serial measurements of left ventricular volume, all of which have been confirmed primarily

TABLE 1. Proposed Criteria for the Diagnosis of Infective Endocarditis Definite Pathologic evidence of vegetations or abscess at surgery or autopsy, or 2 of the following: 1.

Multiple positive blood cultures with no known extracardiac source

2.

Echocardiographic (TTE or TEE) evidence of vegetation, abscess, or new prosthetic valve dehiscence

3.

At least 2 of the following: Fever New or changing murmur Microvascular phenomena such as Osler's nodes or Janeway's lesions

Possible Findings consistent with infective endocarditis but not fulfilling the above criteria

Rejected Firm alternative diagnosis, resolution of illness, or no pathologic evidence of disease after antibiotic therapy for a maximum of 3 days

(Adapted and modified from Lukes A, et al.63 See also: reference 64.)

Texas Heart Institzitejournal


99

| | Suspected I nfective Endocarditis|

Negative

rCinical suspicion

|

Equivocal

Positive

High *

H

High risk *Low risk

Lowl

Observe/look for alternative diagnosis

Negative

|

Positive

|

~~~~~~~~~Appropr-iate antibiotics

TEE in Repeat 7 to IO days

Surgery for

Negative

complications, treatment failure, etc. Look for alternative diagnosis

Fig. 1 Diagnostic approach in patients suspected of having infective endocarditis.

Prosthetic valves, certain congenital defects, persistent bacteremia Virulent organisms; suspicion of perivalvular or significant valvular complications TEE = transesophageal echocardiography; TTE = transthoracic echocardiography (Adapted from Yvorchuk KJ, Chan KL,26 with permission from Mosby-Year Book, Inc.) *

**

with TTE. Except in the most uncomplicated cases, TTE is an essential part of the follow-up of patients

with endocarditis.66f

Uses of TEE in Endocarditis Transesophageal echocardiography is a semi-invasive technique that can be performed quickly and easily in most patients. Imaging of the heart from the esophagus provides a clear view of virtually all cardiac structures that are not obstructed by bones, lungs, or the precordial chest wall. In addition, because of the proximity of the esophagus to the heart, higher transducer frequencies can be applied, leading to improved resolution and more accurate measurement of vegetation size. In particular, small vegetations can be detected more reliably by TEE, as can those attached to prosthetic67 and pulmonary valves.39 100


Daniel and coworkers"' were among the 1st to show the superiority of TEE to TTE for detecting vegetations. In a group of 196 patients strongly suspected of having infective endocarditis, 69 were studied by both TTE and TEE. In 40 of these patients, TEE was performed because of poor images obtained by TTE. Among the 82 echocardiographically abnormal valves found in the 69 patients, TEE demonstrated definite vegetations in 77 of the valves (93%), whereas TTE identified vegetations in only 33 (40%). Other studies comparing TTE and TEE have produced similar results.27324 Lowry and associates'8 found the negative predictive value of TEE to be 100% in native valves and 90% in prosthetic valves. These findings suggest that in patients with native heart valves, a negative TEE examination virtually excludes infective endocarditis. In patients with prosthetic valves, a negative 1,blume Z3, I'Vitinber 2, 1990

TEE result greatly decreases the likelihood of infective endocarditis, but does not completely exclude the possibility.'1 Esophageal tumors, stenoses, diverticula, a perforated viscus, and advanced varices are generally considered to be contraindications for TEE. Previous therapeutic chest irradiation and severe atlantoaxial joint disease (prohibiting sufficient neck flexion) are relative contraindications.69 When conducted by a trained physician, the risk of sequelae associated with TEE is low (10 mm). Those authors' concluded that multiplane TEE is the most accurate technique for the evaluation of valvular or perivalvular lesions in infective endocarditis.

Valvular and Perivalvular Lesions Organisms in a valvular vegetation not only cause destruction of the valve itself, but can also extend beyond the valve and cause perivalvular and myocardial damage.26 This damage may result in valvular regurgitation, a sequela frequently seen in infective endocarditis. A very serious cause of valvular regurgitation is valve leaflet perforation. This condition often has an acute onset, which can lead to marked hemodynamic compromise and death, especially when the aortic valve is involved.78 Valvular insufficiency may also occur when valvular vegetations are so large that they prevent proper coaptation of the valve leaflets (Figs. 2 and 3).79 Perivalvular abscess formation is another extremely important sequela of infective endocarditis for which TEE has repeatedly been shown to Transesophageal Echocardiography in Infective Endocarditis

101

Fig. 2 Biplane transesophageal echocardiogram of a patient with large mitral valve vegetations. Top panel, horizontal plane; bottom panel, vertical plane. LA = left atrium; LV = left ventricle; LVOT = left ventricular outflow tract; MV = mitral valve; V = vegetation

102


Fig. 3 Biplane transesophageal echocardiogram (of patient represented in Fig. 2) showing moderate-to-severe mitral regurgitation. LA = left atrium; LV = left ventricle; LVOT = left ventricular outflow tract; MR = mitral regurgitation; MV = mitral valve; V = vegetation

Volume 23, Number 2, 1996

of cases.8182 These abscesses are usually associated with more severe disease, possibly because antibiotics are unable to penetrate the areas of abscess effectively and the responsible organisms, uninhibited, continue to destroy tissue. Clearly, early diagnosis of endocarditis-related abscesses is crucial. Early surgical therapy may also be extremely important in improving patient outcome.45'83 Daniel and associates45 prospectively examined 118 consecutive patients with infective endocarditis; 44 of those patients had 46 abscess regions confirmed at surgery or autopsy. The aortic valve was most often involved, and staphylococcal organisms were responsible in 52% of the cases. Daniel's group confirmed previously published reports that showed hospital mortality in patients with abscesses to be greater than that of patients without abscesses (23% vs 13.5%). Those authors convincingly demonstrated the superiority of TEE for the detection of endocarditis-related abscesses. The sensitivity and specificity of TTE were 28% and 99%, respectively; whereas the sensitivity and specificity of TEE were 87% and 95%. The difference in sensitivity between TEE and '1'1E was found to be greatest in abscesses involving the interventricular septum and mitral apparatus.45

Fig. 4 Biplane transesophageal echocardiogram of a patient with vegetation of the aortic valve (arrows) and an aortic root abscess. A = abscess; AV = aortic valve; LA = left atrium; RA = right atrium

provide more detailed imaging than TTE (Fig.

4).323536454680 Surgical and autopsy reports have noted perivalvular abscess formations in up to 30% Texas Heart Institutejournal

Evaluation of the Prosthetic Valve Transthoracic imaging of prosthetic valves is extremely difficult because of the acoustic shadowing caused by the metallic parts of the valve. Transesophageal echocardiography is more accurate than 'T'lE for evaluating prosthetic valves, because it provides multiple, close-up views just above and below the valve. This is particularly important in the mitral prosthetic valve, because vegetations usually attach to the atrial side of the valve, an area that is often not well imaged by TTE. For evaluating the aortic valve prosthesis, however, the advantages of TEE over TTE are not so clear.668485 Daniel's group' showed improved sensitivity and specificity of TEE for the evaluation of prosthetic valves. In their study of 148 prosthetic valves in 126 patients, those authors found that the definite diagnosis of prosthetic valve endocarditis was 82% by TEE, compared with 36% by TTE. Mugge and coauthors34 reported similar advantages of evaluating valvular lesions by TEE in a smaller group of patients with prosthetic valves. Clearly, the use of TEE is not necessary in all patients with prosthetic valves and suspected infective endocarditis. Transthoracic echocardiographyspecifically the combination of 2-dimensional and Doppler echocardiography-is a well-established method for accurate hemodynamic studies of prosthetic valves. In addition, 'lITE is very valuable for serial comparisons of valvular lesions in patient follow-up. However, for patients in whom there is a Transesophageal Echocardiography in Infective Endocarditis

103

very strong clinical suspicion of infective endocarditis but negative or equivocal findings by TTE, a complete evaluation with TEE is warranted.

Prognostic Value of TEE in Infective Endocarditis Because echocardiography accurately defines the extent of cardiac involvement in infective endocarditis, many authors have attempted to correlate the presence of vegetations, vegetation size, and mobility with short- and long-term prognoses.2025344~ Recently a meta-analysis was undertaken of 15 published studies on the incidence of endocarditisrelated sequelae in patients with and without echocardiographically detectable vegetations by TTE.86 The study showed that patients with vegetations had an increase in embolic events (from 10% to 37%), severe heart failure (from 26% to 61%), urgent valve replacements (from 12% to 58%), and hospital deaths (from 10% to 23%). Since an abnormal TEE result is equal to (or better than) an abnormal TTE for the detection of valvular vegetations, one can safely assume that vegetations found by TEE indicates a poorer prognosis for the patient. There are also data to suggest that the size and mobility of vegetations may influence the prognosis in patients with infective endocarditis.9 1.20-2- 32.-4 Several TTE studies have repeatedly shown that larger valvular vegetations (length, >10 mm) are more prone to embolize.21-23 Sanfilippo and colleagues25 found that, in addition to size, vegetation mobility, extension, and consistency in 85 patients with left-sided endocarditis were all univariate predictors of outcome in active infective endocarditis.25 Meric and associates8 also reported vegetation mobility to be a highly accurate predictor of embolic events. Considering that TEE has been shown to be more accurate than TTE in defining cardiac involvement in infective endocarditis, Mugge's group34 analyzed whether vegetation size as detected by TEE had any prognostic value in patients with infective endocarditis. In a prospective study of 105 such patients, vegetations larger than 10 mm were associated with a 47% incidence of embolism, compared with a 19% incidence in patients with smaller or no vegetations. When 16 patients who had sustained embolic events before the performance of TEE were excluded, emboli were present in 36% of those with a large vegetation compared with only 6% of patients with smaller or no valvular lesions. This study did not show an increased likelihood of congestive heart failure or death in patients who had the larger valvular vegetations. In our laboratory, we also used TEE to evaluate the prognostic value of vegetation size in patients with active infective endocarditis. Between 1992 and 1995, 40 patients with clinically or histologically 104


proven infective endocarditis and valvular vegetations were identified by TEE. We correlated vegetation sizes with notable clinical events during the hospital stay (Table II) and found that patients with vegetations of 10 mm or greater had a markedly higher incidence of in-hospital events than did those with vegetations of less than 10 mm (92% vs 40%). When the vegetation size was 10 mm or larger, the sensitivity for prediction of in-hospital events was 77%, the specificity was 80%, the positive predictive value was 92%, and the negative predictive value was 53%. Thus, our data confirm the findings of previously published studies that show a strong correlation between the size of valvular vegetations and the incidence of in-hospital complications. TABLE II. Correlation of Degree of Vegetation with Clinical Events in Infective Endocarditis Degree of Vegetation Clinical Factors

210 mm