To develop a method for quantitative analysis of regional left ventricular function from transesophageal two-dimensional echocardiograms, we conducted studies 10 and 20 minutes after induction of anesthesia in 16 patients with normal hearts who were undergoing minor orthopedic operations. Wall thickening was measured with the centerwall method along 100 chords drawn perpendicular to a line constructed around the center of the ventricular wall, midway between the endocardial and epicardial contours. Thickening, either normalized by the length of the end-diastolic perimeter or expressed as a percentage of the end-diastolic wall thickness at each chord, was compared with measurements of endocardial motion. Wall motion was relatively diminished in the anteroseptal region and enhanced on the contralateral wall, but wall thickening was homogeneous throughout the contour. Normalized wall thickening was significantly less variable (standard deviation/mean, 0.47 +/- 0.13) in the normal population than were either percent wall thickening (0.53 +/- 0.012) or wall motion (0.51 +/- 0.09) (p less than 0.005 for both comparisons). There was no significant change in regional or global function between 10 minutes and 20 minutes after the induction of anesthesia. In summary, normalized wall thickening as a parameter of regional left ventricular function is more homogeneous and less variable in subjects with normal hearts than is endocardial motion because wall thickening measurements are not subject to cardiac translocation artifacts. This low variability suggests that normalized wall thickening measured by the centerwall method may prove particularly useful for intraoperative and postoperative monitoring of regional left ventricular function by transesophageal echocardiography in patients undergoing both cardiac and noncardiac surgical procedures.