Background: Although transesophageal echocardiography (TEE) is more sensitive than transthoracic echocardiography (TTE) in detecting echocardiographic evidence of infective endocarditis (IE), the impact of TEE on the clinical diagnosis of IE has not been clearly delineated.
Methods and results: We studied 112 patients with 114 suspected episodes of IE over a 6-year period who underwent both TTE and TEE during their diagnostic evaluation. Using the results of these studies along with clinical and microbiologic data, we attempted to determine the incremental value of TEE to the Duke Endocarditis Diagnostic Criteria. Patients were initially classified into a diagnostic category of the Duke criteria with TTE data, and then the diagnostic classification was reconsidered with TEE data. A diagnostic category reassignment occurred in 25 of 114 episodes of IE evaluated when TEE results were incorporated into the evaluation with the Duke criteria (22 patients were reclassified from possible IE to definite IE whereas 3 patients were reclassified from rejected to possible IE). Diagnostic reclassification occurred in 9 (11%) of the 80 episodes of suspected IE with native cardiac valves and 13 (34%) of 34 episodes with prosthetic cardiac valves. Most patients reclassified from possible IE to definite IE with TEE data (19 of 22) had an intermediate clinical likelihood of IE, whereas 92% of patients had negative TTE results. Pathologic examination of valvular tissue in 22 of the 114 episodes of suspected IE revealed that the positive predictive value of the Duke criteria with TEE data for diagnosis of IE was 85% in patients with native valves and 89% in patients with prosthetic valves.
Conclusions: When clinical evidence of IE is present, TEE improves the sensitivity of the Duke criteria to diagnose definite IE. TEE data appears to be especially useful for the diagnostic evaluation of patients with suspected IE who have prosthetic valves.