[Diagnostic value of echocardiographic examinations in the evaluation of patients with infectious endocarditis of aortic and/or mitral valve localization].

We estimated sensitivity, specificity, predictive value and efficiency of echocardiography in detecting vegetations and ruptured valves in patients with aortic and/or mitral valves infective endocarditis. We studied two groups of patients, in whom both high quality echocardiography examination and surgical inspection of heart valves were available. Group I: 16 patients (32 valves) with aortic and/or mitral valves endocarditis and surgical demonstration of vegetations and/or ruptured valves. In this group the prevalence of vegetations was 65.6%, the prevalence of ruptured valves was 43.7%. Group II was composed of the 16 patients of group I with endocarditis and of 93 other patients without endocarditis. In this group (139 valves) the prevalence of vegetations was 15%, the prevalence of ruptured valves was 13.6%. Echocardiographic detection of valve vegetations. Sensitivity (71.4%) was the same in group I and II. Specificity was 91% in group I and 91.5% in group II. The positive predictive value was 93.7% in group I and 60% in group II. The negative predictive value was 62.5% and 94.7% respectively. Echocardiographic detection of ruptured valves: sensitivity was 50% in I and 42% in group II. Specificity was 94.4% in group I and 99% in group II. The positive predictive value was 87.5% in group I and 89% in group II. The negative predictive value was 70.8% in group I and 91.5% in group II. The echocardiographic efficiency was higher in group II: 88.4% for vegetations and 91.3% for ruptured valves. This reflects the high specificity and the high definition of normal valves in this unselected low-risk group. The efficiency was lower in group I: 78% for vegetations and 75% for ruptured valves. This reflects the poor sensitivity and the failure of the test to identify all the vegetations or the ruptured valves in this selected high-risk group.