Diagnostic precision of echocardiography in mitral valve prolapse.

In order to determine the precision with which currently used echocardiographic criteria can be applied for the diagnosis of mitral valve prolapse, three independent observers (A, B, and C) blindly analyzed the separate M-mode and two-dimensional echocardiograms of 50 patients, 27 of whom had previously been identified clinically as having echocardiographic evidence of prolapse. Observer A's intraobserver repeatability for M-mode echocardiography was 92%, Observer B's was 84%, and Observer C's was 90%. For two-dimensional echocardiography, Observer A's intraobserver repeatability was 98%, Observer B's was 80%, and Observer C's was 82%. We believe that the variability in intraobserver repeatability is related to the frequency with which individual observers diagnosed prolapse. The interobserver repeatability for M-mode echocardiography for Observer A versus B was 64%, for Observer A versus C it was 80%, and for Observer B versus C it was 66%. The interobserver repeatability for two-dimensional echocardiography for Observer A versus B was 54%, for Observer A versus C it was 70%, and for Observer B versus C it was 52%. There was no significant difference between the inter- and intraobserver variability of M-mode versus two-dimensional echocardiography. Review of cases in which readings were discrepant revealed that these cases usually had relatively mild changes. Clinicians should be aware of the inherent variability of echocardiographic interpretation when they make a diagnosis of mitral valve prolapse.