Reproducibility of quantitative myocardial contrast echocardiography.

To determine whether myocardial contrast echocardiography is quantitatively reproducible, repeated intracoronary injections of sonicated albumin (5%) were performed in eight open chest dogs. Paired injections were performed at baseline, during ischemia produced by ligation of a coronary artery, and during hyperemia induced by intravenous infusion of 0.75 mg/kg body weight of dipyridamole. Contrast washout curves were generated for the left anterior descending coronary artery territory (ischemic area) and left circumflex coronary artery territory (nonischemic area) by beat per beat analysis of frozen end-diastolic frames of left ventricular short-axis views. Peak contrast intensity, contrast washout half-time and area under the curve were derived from these curves. A total of 75 contrast washout curves were analyzed for the study of interinjection, intraobserver and interobserver reproducibility. The correlation coefficients between measurements obtained from paired injections of the echocardiographic contrast agent (interinjection reproducibility) ranged from 0.78 for peak contrast intensity to 0.87 for area under the curve. Percent error varied between 14.7% and 24.7%. The intraobserver variability in measurements was less than the interinjection variability, with a cumulative mean percent error of 17.8% and correlation coefficients of 0.72 (peak contrast intensity), 0.95 (area under the curve) and 0.96 (washout half-time). Interobserver correlation for all indexes was high (r = 0.92 to 0.96). It is concluded that peak contrast intensity, contrast washout half-time and the area under the curve derived from myocardial contrast washout curves can be measured reproducibly from videotapes. In addition, the variability between two injections attempted under identical conditions is greater than reader variability from videotapes.