VALUE OF TRANSTHORACIC ECHOCARDIOGRAPHY IN PREDICTING EMBOLIC EVENTS IN ACTIVE INFECTIVE ENDOCARDITIS

The purpose of this study was twofold: (1) to determine interobserver variability of echocardiographic characteristics of vegetations in patients with infective endocarditis, and (2) to assess the value of these vegetation characteristics in predicting embolic events. Although echocardiography contributes to the diagnosis of patients with infective endocarditis, its prognostic role in predicting embolic events is controversial. The echocardiograms of 41 patients with infective endocarditis were independently reviewed by 4 echocardiographers blinded to the clinical data. If a vegetation was present, the following char acteristics were analyzed: involved site, size, mobility, shape, and pedunculated or sessile attachment. Each echocardiographer also made a ''gestalt'' estimate of embolic risk based on these vegetation characteristics. Interobserver agreement on vegetation characteristics and their relation to embolic events was then determined using kappa statistics and logistic regression analysis. Interobserver agreement was 98% with regard to echocardiographic vegetation presence and 97% with regard to the involved site. Of the 30 patients in whom vegetations were observed, complete observer agreement was achieved with regard to size in 22 (73%), mobility in 17 (57%), shape in 11 (37%), and attachment in 12 (40%). Vegetations with a maximal diameter of > 10 mm were associated with a 50% incidence of embolic events, compared with a 42% incidence of emboli in patients with vegetations measuring less than or equal to 10 mm. Interobserver variability was great with respect to vegetation shape, mobility, and attachment characteristics. Echocardiographic vegetation characteristics were not helpful in defining the risk of embolic complications in patients with endocarditis.