My Cardiologist explained that the infection in the Aorta was so advanced it took all of the surgeon’s skills just to be able to attach the valve in place. —Lewis Grizzard

This quotation is from a talented journalist who, had he survived the ravages of endocarditis, might have gone on to be recognized as the world’s foremost commentator and humorist. But sadly, in 1994, he succumbed to the disease, dying at an extremely young age. His case highlights the fact that infectious endocarditis (IE), with an annual incidence of 20,000 to 30,000 new cases, continues to be a devastating illness with high morbidity and mortality. Compared with the advances that have occurred in the past 20 to 30 years in the prevention, diagnosis, and treatment of coronary artery disease and congestive heart failure, to name a few of the conditions that we cardiovascular specialists deal with, IE stands out, because despite dramatic developments in imaging and treatment, this condition still has strikingly high morbidity and mortality. And why is that the case? In my opinion, it is due in part to the marked variability in the clinical presentation of endocarditis: it can mimic inflammatory conditions or malignancies, and its symptoms can be masked in the elderly, leading to delays in diagnosis. Also, I believe that despite important advances in imaging, there is often a delay in instituting appropriate interventions early in the course, interventions that might prevent the catastrophic valvular and perivalvular consequences of IE.

With this background, the article by Kini et al in the current issue of JASE does give us a chance to revisit an important question: what are the appropriate roles of transthoracic echocardiography (TTE) and transesophageal echocardiography (TEE) in the diagnosis and potentially in the management of IE? Some of the findings of Kini et al are neither unique nor new (eg, the finding that TEE has improved diagnostic accuracy for detecting valvular vegetations), but other findings (such as the lack of a correlation of noxious organisms such as Staphylococcus aureus or fungal infections with valvular vegetations detected on TEE) are counter to current literature and clinical experience.

Kini et al present their findings from a retrospective review of nearly 3000 echocardiographic studies done between 2003 and 2008 in patients examined because of suspicion of IE. Among these 3000 studies, the authors evaluated slightly more than 500 consecutive pairs of transthoracic and transesophageal studies, all done within 1 week of each other. Obviously, this study had a major selection bias, in that nearly 2500 transthoracic echocardiographic exams were performed for the indication of known or suspected endocarditis, but the findings did not lead to further evaluation by TEE. Although the authors recognize this fact, they do not make clear whether this was because the findings on TTE were so clearly indicative of valvular vegetations that TEE was not deemed necessary or, as seems more likely, the primary indications for initial TTE were marginal (eg, 1 episode of fever in an individual with negative blood cultures), and clinicians believed that no further testing was necessary after transthoracic evaluation was negative. Hence, in essence, this paper really is not able to answer the question of when TTE and/or TEE is indicated in patients with known or suspected endocarditis. The authors do, however, indirectly offer suggestions as to how that question might be answered.

This brings us to the question of whether guideline documents or published papers answer this question. Two stimulating articles, one an editorial by Shaneyfelt and Centor and the other an article by Tricoci et al, address the facts surrounding the American College of Cardiology and American Heart Association clinical guidelines. Both papers highlight the fact that guidelines are, in essence, based on expert opinion, a finding I do not believe necessarily minimizes the importance of guidelines or consensus statements. Experts, especially if they are clinicians who have had extensive experience with the use of diagnostic and therapeutic modalities in managing multiple clinical conditions, can give direction as to how diagnostic and therapeutic interventions can best be used. To this point, two recent outstanding articles in the echocardiographic literature highlight the question of which echocardiographic diagnostic modality should used in evaluating patients with suspected IE.

Douglas et al published appropriateness criteria for TTE and TEE in 2007 and suggested that TEE is the appropriate initial test in patients with moderate to high pretest probability for endocarditis, especially those with suspected S aureus infections or fungal infections. These authors highlighted an extremely important point: TEE not only serves a diagnostic role but could play a prognostic role by detecting complications of endocarditis, which so often lead to serious morbidity or mortality. Also, recently, Zoghbi et al published an outstanding article on the use of echocardiography for the evaluation of prosthetic valves. When examining the question of prosthetic valve endocarditis, these authors again emphasized that TEE, because of its high sensitivity and specificity for detecting valvular vegetations and complications, should be considered the initial test in patients with prosthetic valves who have medium to high clinical index of suspicion for endocarditis. They also pointed out that no echocardiographic technique should be considered part of a routine “fever screen.” These 2 outstanding articles do give us some direction, but let us consider other points.

We know that TEE has improved diagnostic accuracy for detecting both native and prosthetic valve endocarditis, with TTE detecting vegetations in proven native valve endocarditis in only 50% to 60% of cases and in as few as 20% of cases of prosthetic valve endocarditis. Does this mean that a well-performed transthoracic echocardiographic study with cardiac Doppler has no diagnostic or therapeutic bearing on patients with suspected implications in the setting of native valve endocarditis? Not necessarily, because carefully performed and interpreted TTE in the setting of positive blood cultures and native valvular anatomy may detect sizable valvular vegetations, valvular dehiscence, and even sizable abscess and fistula formation, as well as provide important hemodynamic information. There is a caveat here, though, because in today’s environment, we clinicians see many patients who have had prolonged hospitalization with multiple insertions of right-sided lines, and those who may have acquired nosocomial infections. We also see patients who have indwelling catheters used for the treatment of oncologic conditions, as well as those who have pacemakers and implantable cardioverter-defibrillators in the right-sided veins and right-sided cardiac chambers. In these individuals, negative results on TTE do not necessarily exclude the possibility of vegetative lesions that could reside on catheters in the superior vena cava (SVC), pannuses or thrombi in the right atrium, or small masses attached to the pacer leads. Hence, in individuals who have indwelling lines or who have recently been hospitalized, and in whom there are positive blood cultures or a high index of suspicion for endocarditis, TEE may be extremely valuable for interrogating the SVC, the right atrium, and pacemaker or implantable cardioverter-defibrillator leads to rule out the possibility of an infected pannus or thrombus.

Kini et al suggest that increasing severity of valvular regurgitations in patients with suspected IE could serve as an indication for follow-up TEE and for the detection of valvular vegetations. Undoubtedly, native and prosthetic valve endocarditic lesions can produce ongoing valvular and perivalvular destruction, leading to increasing grades of regurgitation. Clearly, if detected, this may be a sign of ongoing valvular destruction and may warrant further workup with TEE. However, Kini et al did not quantitate the severity of regurgitant lesions on serial transthoracic echocardiographic studies, and this is problematic from both an interpreter’s perspective and a performance standpoint. Obviously, different interpreters could interpret regurgitant lesions differently, and alterations in hemodynamic conditions, such as blood pressure, clearly can affect findings on serial imaging of regurgitant lesions. Saying this, I do believe that increasing “severity” of regurgitant lesions could be a strong indicator of ongoing valvular destruction or even abscess or fistula formation and should warrant TEE to evaluate native and prosthetic valves more carefully for endocarditic lesions.

TEE clearly plays a critical role in those with known or suspected endocarditis, especially prosthetic endocarditis, because of its ability to detect complications of infection, such as abscess formation and fistula formation. Numerous authors, including Daniel et al, Cosmi et al, and Graupner et al, have highlighted the diagnostic and prognostic value of TEE in detecting periannular extension and abscess formation in left-sided endocarditic lesions. Graupner et al showed that TEE detected periannular extension in 37% of nearly 211 patients with established left-sided endocarditis. The majority of these cases (55%) involved prosthetic valves, and interestingly, those with coagulase-negative staphylococcal infections or previous endocarditis were more likely to have such extension.

Recently, Cosmi et al substantiated the devastating consequences of perivalvular abscesses, highlighting the importance of timely diagnosis. Hence, I do believe that TEE may well be the initial diagnostic procedure of choice in presumed aortic valve endocarditis, whether native or prosthetic, because of its ability not only to detect valvular vegetation and document the hemodynamic consequences, but also, and more important, to document periannular abscess formation or fistula extension into the right atrium, right ventricular outflow tract, or even through the mitral-aortic intervalvular fibrosa. Not only is the diagnosis important, but findings on TEE often have major therapeutic implications, as they can appropriately direct proper surgical intervention, aiding the surgeon in knowing beforehand the complications that he or she might encounter in the operating room.

A striking finding of Kini et al, which goes against a great deal of clinical and published experience, was their lack of correlation of S aureus bacteremia or fungemia with the eventual finding of vegetations in their echocardiographic studies. Clearly, the literature is replete with numerous studies suggesting that S aureus bacteremia and fungemia are not only associated with increasing likelihood of sizable valvular vegetations but also suggestive of very aggressive infectious agents, which can produce marked valvular and perivalvular tissue destruction. Many infectious disease experts, as well as cardiologists and cardiac surgeons, believe that the type of infecting organism is not just of clinical importance: it also influences the pretest likelihood of detecting not only a vegetation but also the above-mentioned complications. Hence, in the setting of suspected or known S aureus infection or fungal infection, I believe that TEE should be the imaging procedure of choice.

Fungal endocarditis is said to cause 1% to 10% of all endocarditic lesions and could occur in up to 15% of all prosthetic valve endocarditis, although this has not been my own experience. It is important to remember that fungal lesions can occur not only on prosthetic valves but also on endovascular grafting tissue. Fungal endocarditis carries very high morbidity and mortality, often due to delays in diagnosis, as well as the fact that large vegetative lesions cause extensive tissue destruction and often produce extensive systemic embolization. The most common infecting fungal organism is Candida . It is important for the reader to remember that there are major risk factors for the development of fungal endocarditis, and individuals who have these risk factors and suspicion of fungal endocarditis should undergo TEE, because of its ability to detect these vegetative lesions not only on valves but also on prosthetic material, such as Bentall grafts and other vascular grafts. The major risk factors for the development of fungal endocarditis are (1) prior cardiac surgical procedures, especially the insertion of prosthetic valves, prosthetic devices, or endovascular grafts; (2) prolonged hospitalization in intensive care units, often with multiple indwelling central venous catheters; and (3) the use of long-term antibiotics.

Kini et al do raise the awareness of culture-negative endocarditis. However, I found their statements somewhat confusing: they found that negative blood cultures were not associated with the finding of vegetative lesions on subsequent TEE, and yet they point out that 14% of their patients with vegetations seen on TEE, including nearly a third of those with prosthetic valve vegetative lesions, had culture-negative or nonbacterial thrombotic endocarditis. It’s important for the reader to remember that culture-negative endocarditis constitutes approximately 5% to 10% of all cases of endocarditis and may be an increasingly serious problem. That is because the endpoints of death and/or surgery have been said to occur more frequently in culture-negative endocarditis. Part of the difficulty in this diagnosis is that this condition may occur in individuals who have been partially or inappropriately treated with antibiotics, hence masking the ability to culture the offending organism. Also, infections with fastidious organisms, such as HACEK ( Haemophilus spp, Actinobacillus actinomycetemcomitans , Cardiobacterium hominis , Eikenella corrodens , and Kingella ), Bartonelli spp (especially in homeless men with substandard hygiene), and Coxiella burnetti , and even fungal infections, all should be considered in patients in whom there is suspicion of IE and yet cultures are negative. Again, TEE is of extreme importance in detecting vegetative lesions on both native and prosthetic valves, as well as on prosthetic material or in great vessels, such as the SVC or the aorta, because of its anatomic superiority to image these sites, as well as enhanced resolution.

Although not directly highlighted by Kini et al, nosocomial IE is being recognized increasingly, often within weeks or months after prolonged hospitalization or hospital-based procedures. Nosocomial IE, especially with S aureus , can affect right-sided structures of the heart because of the placement of intravascular lines (in the vena cava) or procedures (insertion of pacemakers, etc), which could possibly damage the endothelial surface of the SVC, the right atrium, or the right-sided valves. The second most common type of nosocomial pathogens associated with endocarditic lesions, which often affect left-sided valves, is seen in patients who have had recurrent or multiple genitourinary manipulations and instrumentation and then develop infections caused by Enterococci . As with culture negative endocarditis, when there is suspicion of infections on pannuses, thrombi, or even devices located in the SVC, right atrium, or right ventricle, TEE could be considered as the procedure of choice.

Kini et al are to be complimented for giving the readers a chance to reflect not only on their findings but also on current thinking about the appropriateness of echocardiographic techniques in patients with known or suspected IE. I have been fortunate to spend nearly 34 years as a clinician intimately involved in the clinical use of echocardiography to diagnose and manage various cardiovascular conditions. It has been just as long since I first had the opportunity to become involved in the early field of 2-dimensional TTE, an opportunity afforded me while I was a fellow and junior faculty member at Stanford Medical Center under the mentorship of one of the true giants in the field, Dr Richard Popp. From our earliest days with 2-dimensional TTE at Stanford, it was apparent that this echocardiographic technique was unique in its ability to evaluate patients with known and suspected native and prosthetic valve endocarditis. Over the ensuing years, I have maintained a strong interest in and been a proponent of the appropriate use of both TTE and TEE in patients with known or suspected IE. The key phrase here is “appropriate use.” No echocardiographic technique and in fact no imaging technique should be practiced or used in a vacuum or believed to be 100% accurate. Excellence in clinical judgment, and excellence in the performance and interpretation of these echocardiographic studies, is critical to providing optimal patient care.