Variability in the measurement of regional left ventricular wall motion from contrast angiograms

Four types of variability affecting quantification of regional wall motion from contrast left ventriculograms (LVgrams) were studied. These included beat-to-beat variability in 24 LVgrams, intraobserver and interobserver variability in 20 LVgrams, and study-to-study variability in serial LVgrams of 21 patients with stable coronary artery disease. Motion was measured at 100 equidistant chords perpendicular to a center line drawn midway between the end-diastolic and end-systolic contours and normalized for heart size. Variability was computed as the absolute difference between observations. Beat-to-beat, intraobserver, and interobserver variability at the 100 chords were similar, averaging 14%, 14%, and 17%, respectively, of the mean motion in 64 patients with normal ventriculograms. Study-to-study variability was significantly higher, averaging 30% of mean normal motion, but was reduced when regional motion was calculated as the mean motion of chords within a region of interest. Variability peaked at the apex. Realignment to correct for cardiac rotation significantly increased variability. Investigators whose methods of wall motion analysis rely on identification of the apex as a landmark should be aware of this source of potential variability and error.