Diagnostic accuracy of transesophageal echocardiography for detecting left atrial thrombi in patients with rheumatic heart disease having undergone mitral valve operations

Abstract Transesophageal echocardiography (TEE) has emerged as an efficient method for detecting left atrial (LA) thrombi in recent years, but its accuracy has not been fully evaluated. A prospective clinicopathologic study in 213 consecutive patients with chronic rheumatic mitral valve disease over a period of 39 months was undertaken. All patients underwent open heart surgery within 3 days after the TEE study. The presence or absence of LA thrombi was confirmed at surgery by direct inspection of the left atrium and proven by histopathologic examination. Of the 213 patients, 147 had predominant mitral stenosis, and the remaining 66 patients had significant mitral regurgitation. Twenty-eight patients had LA thrombi by TEE criteria. These findings were all confirmed at surgicopathologic studies (specificity 100%). However, in 2 patients, LA thrombi were present but could not be detected by TEE (sensitivity 93.3%). Therefore, the positive predictive value was 100%, the negative predictive value was 98.9% and the diagnostic accuracy was 99.1%. No thrombi were found in patients with significant mitral regurgitation. The frequency of LA thrombi in patients with predominant mitral stenosis was 20% (30 of 147), and most of these patients had chronic atrial fibrillation (28 of 30, 93%). Only 16 patients (16 of 30, 53%) were found to have LA thrombi by transthoracic echocardiography. Furthermore, our data showed poor correlation between the echogenicity of LA thrombi and the degree of thrombus organization. Thus, TEE is excellent for detecting LA thrombi in patients with rheumatic heart disease severe enough to warrant mitral valve operations.