Current issues of ACP Journal Club are published in Annals of Internal Medicine


Transesophageal echocardiography was more sensitive than transthoracic echocardiography in detecting endocarditis-associated abscesses

ACP J Club. 1991 July-Aug;115:21. doi:10.7326/ACPJC-1991-115-1-021

Related Content in the Archives
Two-dimensional echocardiography for perivalvular abscesses

Source Citation

Daniel WG, Mügge A, Martin RP, et al. Improvement in the diagnosis of abscesses associated with endocarditis by transesophageal echocardiography. N Engl J Med. 1991 Mar 21;324:795-800. [PubMed ID: 1997851]



To determine whether transesophageal echocardiography is superior to transthoracic echocardiography in detecting abscesses in patients with infective endocarditis.


Comparison of transesophageal and transthoracic echocardiographic findings with unblinded findings at surgery or autopsy.


Medical and surgical service of a German medical school.


Consecutive series of 118 patients from 1984 to 1989, with infective endocarditis who, before going to surgery or autopsy, had transthoracic and transesophageal echocardiograms. During this period, 28 similar patients survived on medical therapy alone (all had negative echocardiograms for abscess) and 2 other patients with severe aortic regurgitation and shock went immediately to surgery without preoperative echocardiography (neither had an abscess).

Description of test and diagnostic standard

The esophageal transducer was mounted on a modified gastroscope and passed into fasting patients in the supine left lateral position; the procedure usually was done within 5 minutes. Transthoracic M-mode and 2-dimensional echocardiograms were done in the usual way. Each was read by 2 independent observers, who resolved disagreements by consensus. The diagnostic standard was direct inspection at surgery or autopsy and histopathologic examination. Surgery was done only for heart failure or persistent infection, not for echocardiographic findings of abscess. On echocardiography, definite regions of reduced echodensity and echolucent cavities within or adjacent to the valvular annulus were considered abscesses. At surgery or autopsy, any region of necrosis containing purulent material and penetrating into the myocardium or valvular annulus was considered an abscess.

Main outcome measures

Sensitivity and specificity.

Main results

44 (37%) of the 118 patients who had surgery or an autopsy had 46 abscesses. Transthoracic echocardiography detected only 13 of these 46 abscesses, but transesophageal echocardiography detected these 13 plus 27 others (P for the difference < 0.001). The sensitivity, specificity, and likelihood ratio for a positive and negative test result for transthoracic echocardiography were 28%, 99%, {28, and 0.7}* and for transesophageal echocardiography were 87%, 95%, {17, and 0.14}*, respectively. The 2 observers disagreed for 3.4% of transthoracic echocardiograms and 4.2% of transesophageal echocardiograms.


In the detection of abscesses in patients with infective endocarditis, transesophageal echocardiography was much more sensitive than, and about as specific as, transthoracic echocardiography.

Source of funding: Not stated.

Address for article reprint: Dr. W.G. Daniel, Division of Cardiology, Department of Internal Medicine, Hannover Medical School, Konstanty-Gutschow-Strasse 8, D-3000 Hannover 61, Germany.

*Numbers calculated from data in article.


Before antibiotics were discovered, infective endocarditis was almost uniformly fatal. Although the prognosis is now markedly improved, it remains a disease with considerable mortality and morbidity. Patients with abscesses are much less likely to be cured with medical therapy alone. Thus, the early detection of abscesses might be useful in leading to surgery before complications, such as congestive heart failure, develop. Transthoracic echocardiography is the standard method for diagnosing vegetations, but it is not very effective in diagnosing abscesses. Recently, transesophageal echocardiography has been documented to improve the diagnosis of vegetations. On the basis of this study, transesopha- geal echocardiography is also useful for diagnosing abscesses.

Although the results of this study are acceptable on the basis of the sensitivity and specificity, the predictive values reported must be adjusted for local circumstances. As noted by the authors, the population studied was skewed. First, most patients were referred from other hospitals, so presumably they were sicker. Second, they were all patients who had had surgery or who had died. As such, the proportion of these patients with an abscess would be much higher than the proportion of all patients with endocarditis. Third, the surgeons' awareness of the echocardiographic findings may have increased intraoperative detection of abscesses.

These limitations could not have been eliminated. Only by requiring histologic evidence is the diagnosis of abscess guaranteed. In addition, it would be unethical to withhold the results of the echocardiograms from the surgeons. Regardless, the results of this study show that transesophageal echocardio- graphy is a useful diagnostic tool for searching for abscesses in patients with infective endocarditis who do not respond to adequate medical therapy.

Brian Strom, MD
University of PennsylvaniaPhiladelphia, Pennsylvania, USA