The effect of high thoracic epidural anesthesia (TEA on left ventricular contractility was studied in a prospective clinical trial. Forty-eight patients with ASA physical status 1 and 2 and without cardiovascular disease were included in the study. Thirty-six patients scheduled for elective upper abdominal surgery were randomly assigned to Group 1 (TEA, bupivacaine 0.25%, n = 12), Group 2 (TEA, bupivacaine 0.5%, n = 12) or to Group 3 (control without TEA, n = 12). TEA induced a sensory block which extended over all cardiac segments. In order to assess the effect of systemically absorbed bupivacaine, we studied a separate group of patients who received lumbar epidural anesthesia without involvement of the cardiac segments: Group 4 (LEA, bupivacaine 0.5%, n = 10). Left ventricular contractility was assessed using the end-systolic pressure-length relationship. Left ventricular dimensions were measured by transesophageal echocardiography. All hemodynamic measurements were performed under general anesthesia. There was no significant difference in systolic or diastolic arterial pressure, heart rate, left ventricular end-systolic and end-diastolic cross-sectional areas and left ventricular wall stress between the four groups. Left ventricular maximum elastance as a measure of left ventricular contractility was significantly (P < 0.001) reduced in Groups 1 and 2 [8.1 (+/- 3.5) and 9.6 (+/- 4.4) kPa.cm-1, respectively] as compared to Groups 3 and 4 [18.4 (+/- 8.8) and 17.7 (+/- 7.7) kPa.cm-1, respectively]. No significant difference could be demonstrated between Groups 1 and 2 or between Groups 3 and 4. It is concluded that high TEA severely alters left ventricular contractility even in subjects without pre-existing cardiac disease.