Analysis of valvar and left ventricular parameters in infective aortic endocarditis as predictors of outcome: a combined assessment by transesophageal and strain echocardiography

Purpose. The timing of surgery is crucial for patients with aortic endocarditis in whom medical therapy fails. The aim of our study is to identify potential echocardiographic "markers" of adverse events in patients with aortic regurgitation from infective endocarditis. Methods. Fifteen patients with aortic regurgitation (AR) from infective endocarditis were studied by transesopageal echocardiography (TEE) and transthoracic speckle tracking echocardiography (STE). Fifteen healthy subjects were selected as controls. Vegetation size was assessed by TEE. Standard transthoracic echocardiographic parameters were determined. Global left ventricular (LV) longitudinal strain (LS), radial and circumferential strain were measured by STE. Averaged LV rotation and rotational velocities from the base and apex were obtained and used for calculation of LV torsion (LVtor). Mitral annular velocities were also obtained by tissue Doppler imaging (TDI). Results. Mean percentage intraobserver variability was 6% for LV-LS and 8% for LV-tor, and mean percentage interobserver variability was 11% for for LV-LS and 12% for LV-tor. Severe AR had decreased LS compared with control subjects. LVtor decreased significantly in severe AR compared to normals (p<.001) as a result of a predominant decrease in apical rotation. By multivariate analysis, LV-LS (p=0.03), LV-tor (p=0.008) and vegetation size (p=0.009) were predictive of adverse events. ROC curves suggested that thresholds offering an adequate compromise between sensitivity and specificity for adverse events detection were -18.4% for mean global LV-LS (AUC .76), 12mm for vegetation size (AUC .84), and 19.7degrees for LVtor (AUC .89). The combination of vegetation size and LVtor had the highest diagnostic accuracy for identifying adverse outcome, superior to vegetation size (p=.008) or LVtor alone (p=.026). Conclusions. The combined evaluation of the characteristics of vegetating masses and LV function strain parameters improve the sensitivity of echocardiographic indices in predicting cardiac morbidity and mortality of aortic regurgitation from infective endocarditis.