Transesophageal echocardiography in diagnosis of infective endocarditis.

STUDY OBJECTIVE: To determine whether transesophageal echocardiography (TEE) was superior to transthoracic echocardiography (TTE) in defining valvular vegetations and diagnosing clinical infective endocarditis (IE) in patients suspected of having this infection.

PATIENTS AND METHODS: Between April 1989 and May 1991, 64 febrile patients with clinical and/or microbiologic risk factors for IE were prospectively enrolled. Patients underwent both TEE and TTE, which were interpreted in a blinded fashion as to the patient's clinical status. Clinical criteria for the diagnosis of IE were compared with TEE and TTE findings to delineate the ability of the two echocardiographic techniques to define valvular vegetations and to establish the clinical diagnosis of vegetative IE.

RESULTS: Thirty-four valves had typical valvular vegetations demonstrated by either TEE or TTE. Transesophageal echocardiography was more sensitive than TTE in identifying valvular vegetations (33/34 vs 23/34 instances, respectively; p = 0.004). Also, TEE was better at identifying smaller vegetations (< 1 cm) than TTE; 12 patients with such vegetations were identified by TEE as compared with only 5 of 12 identified by TTE (p = 0.02). Of the 64 patients enrolled, 30 (47 percent) were classified as having "definite" or "probable" IE by modified von Reyn criteria. Among these 30 patients, TEE was significantly more sensitive than TTE at documenting vegetative valvular lesions (26/30 [87 percent] vs 18/30 [60 percent], respectively) (p < 0.01). Both TEE and TTE were highly specific (91 percent) in delineating valvular vegetations in this patient population; two of the three false-positive TEE studies for valvular vegetations occurred in patients with a history of IE. All nine periannular complications of IE were identified by TEE, as compared with only two being defined by TTE (p = 0.001).

CONCLUSIONS: Transesophageal echocardiography is significantly more sensitive than TTE and highly specific in both confirming the clinical diagnosis of IE, as well as in identifying valvular vegetations in patients at risk for this infection. Our data also support the concept that TEE is the echocardiographic method of choice for defining small vegetations and periannular complications in IE.