Infective endocarditis remains an all too common illness, with a significant risk of morbidity and mortality (Chap. 86). Traditionally, the diagnosis has been based on either the cumulative results of blood cultures, physical examination, and laboratory findings or on pathologic proof of infected valvular vegetations at surgery or autopsy. Echocardiography may play an important role in infective endocarditis in regard to diagnosis, detection of associated cardiac abnormalities and hemodynamic dysfunction, prognosis, and the need for surgery. Vegetations can now be visualized noninvasively in many (but not all) cases of endocarditis and have become the echocardiographic hallmark of this disorder.198,199 Thus even though TTE cannot exclude endocarditis, abnormal findings may strongly suggest the disorder, even in the presence of negative blood cultures. Since no single abnormality has 100 percent diagnostic accuracy for infective endocarditis, strategies for diagnosis have been devised based upon a number of criteria,200 and definite echocardiographic vegetations are designated as a major criterion. Both TTE and TEE are valuable in the detection of perivalvular abscesses and prosthetic-valve endocarditis.102 Although there is considerable debate concerning the most accurate diagnostic criteria for endocarditis, echocardiography has become one of the most commonly used techniques for the evaluation of potentially affected patients.201 Echocardiography (both TTE and TEE) is also useful for evaluation of patients with systemic lupus erythematosus complicated by Libman-Sacks endocarditis.202 Even though M-mode recordings produced the first echocardiographic description of vegetations, this modality has been largely replaced by 2D imaging. With 2D echocardiography, valvular vegetations typically appear as irregular, usually localized masses of varying echocardiographic density attached to valvular or perivalvular structures (Figs. 15–94 and 15–95) without significantly altering their mobility. The vegetations may be small or quite large and may attach directly to the valve leaflets or the supporting chordal apparatus.198,199,203 Occasionally, vegetations may be attached to unusual structures, such as the atrial wall or the eustachian valve.204,205 Both small, nonmobile vegetations on a normal valve and large vegetations on a markedly abnormal valve may be difficult or impossible to identify with certainty. Aggressive infections often cause perforation or distortion of the affected leaflet, leading to varying degrees of valvular regurgitation. This is distinctly different from most cases of nonbacterial thrombotic (marantic) endocarditis, where the valvular vegetations are usually nondestructive. In cases of infective endocarditis, the presence of vegetations by TTE increases the risk of heart failure, embolic events, and the ultimate necessity of valve replacement.206 Unfortunately, TTE is not 100 percent sensitive in detecting vegetations, and up to 20 percent of patients with proved native-valve endocarditis may have unremarkable examinations. The sensitivity of TTE in prosthetic valve endocarditis has been found to be even lower (approximately 60 percent) due to technical limitations in imaging. | | Figure 15–94. A. Apical four-chamber view demonstrating a large tricuspid valve vegetation (arrow). RA = right atrium; LA = left atrium; LV = left ventricle; RV = right ventricle. B. Parasternal long axis view demonstrating a vegetation (arrow) on the anterior valve leaflet; AO = aorta. |
| | Figure 15–95. Longitudinal TEE view of a large mitral valve vegetation (arrow). a = left atrium; v = left ventricle. (Courtesy of William D. Keen, Jr., MD.) |
TEE has proved significantly more sensitive than TTE for detection of infective vegetations and is extremely helpful for the diagnosis of perivalvular abscesses, mycotic diverticula, and prosthetic valve involvement.102,207 The technique is also useful for assessing valvular regurgitation, fistulas (Fig. 15–96), other hemodynamic complications of endocarditis, and risk of embolization.208 Although a negative TEE examination cannot completely exclude infective endocarditis, it confers a relatively good prognosis in those cases where the diagnosis is eventually confirmed. The optimal use of TEE in suspected endocarditis remains controversial: some authorities recommend routine TEE in all cases, but many do not. A reasonable approach may be to perform TTE as the first screening test in patients with suspected endocarditis. If the study is technically difficult or equivocal or detects vegetations in patients at high risk for perivalvular complications or hemodynamic compromise, TEE should be performed. If TTE is unremarkable or detects vegetations in patients at low risk for complications, TEE may not be necessary.103 Exceptions to this last recommendation might include patients with prior antibiotic treatment or those with persistent bacteremia or fever of unknown etiology. In high-risk patients (i.e., with possible prosthetic valve involvement, congenital heart disease, or infection with especially virulent organisms), TEE is recommended even if TTE is normal.103 | | Figure 15–96. Longitudinal TEE image demonstrating a fistula between the aorta (A) and left atrium (LA) in a patient with endocarditis. AV = aortic valve; P = pulmonary artery; LV = left ventricle; M = mitral valve. (From Sobel J, Maisel AS, Tarazi R, Blanchard DG. Gonococcal endocarditis: Assessment by transesophageal echocardiography. J Am Soc Echocardiogr 1997;10:367–370. With permission.) |
Echocardiographic evaluation of suspected endocarditis is not without pitfalls. It may be quite difficult to detect active vegetations in patients with preexisting valvular abnormalities such as calcification, myxomatous change, rheumatic involvement, and healed vegetations. Despite recent technologic advances, the diagnosis of infective endocarditis remains a clinical one, and overreliance on echocardiography may cause mistakes. Therefore, echocardiographic results should be integrated with other clinical information to diagnose this disorder accurately.209 |