ORIGINAL ARTICLES A zyxwvu Comparison of Methods for the Detection of Myocardial Ischemia During Noncardiac Surgery: Automated ST-Segment Analysis Systems, Electrocardiography, and Transesophageal Echocardiography John E. Ellis, MD, Manish N. Shah, BA, Joan E. Briller, MD, Michael F. Roizen, MD, Solomon Aronson, MD, and Steven B. Feinstein, MD Department of Anesthesia and Critical Care and Department of Internal Medicine, The University of Chicago, Chicago, Illinois Clinicians often fail to detect intraoperative ischemic electrocardiographic (ECG) changes when viewing oscilloscopes. Automated ST-segment monitors promise to increase the detection of such ECG changes. We investigated the capacity of two com- mercially available ST-segment monitors to detect intraoperative myocardial ischemia in patients at high risk for developing intraoperative myocardial isch- emia during vascular and other noncardiac proce- dures. The ST-segment monitors were compared with two reference monitors: (a) printed eight-lead ECGs, as interpreted by a cardiologist, and zyxwvu (b) the presence of segmental wall motion abnormalities and thickening abnormalities detected by transesoph- ageal echocardiography (TEE). We also examined the capacity of the printed ECG to diagnose myocardial ischemia when compared with TEE. We studied zyxwvu 44 patients who underwent TEE, printed multilead ECG, oscilloscope monitoring of leads V, and 11, and measurement of ST-segment deviation from the base- line using an automated Hewlett Packard ST-segment device. The sensitivities for the Hewlett Packard system were 40% for TEE-diagnosed myocardial ischemia and 75% for ECG-diagnosed ischemia. Comparison of the printed ECG with TEE revealed that ST-segment changes in the printed ECG, as analyzed by a cardiologist, were 25% sensitive and 62% specific for the detection of TEE-diagnosed myo- cardial ischemia. When T-wave inversions were added to ST-segment depression as a criterion for the diagnosis of myocardial ischemia by the printed ECG, the sensitivity of ECG for the detection of intraoperative myocardial ischemia, as determined by TEE, was 40% and specificity was 58%. Twenty- three of the 44 patients were simultaneously moni- tored in leads I, 11 , and V, with an automated Marquette ST-segment monitor. In the 23 patients monitored with both Hewlett Packard and Marquette systems, the sensitivities (80% vs 100%, respectively) and specificities (67% vs 50%, respectively) were similar for ECG-diagnosed myocardial ischemia. Monitors of intraoperative myocardial ischemia often do not agree with each other; however, auto- mated ST-segment monitors predict most ischemic changes seen on the printed ECG and can be used as an alarm to alert the clinician to examine the ECG. (Anesth Analg 1992;75:76&72) ecent work has documented the high incidence of perioperative ischemia in patients with pe- known coronary artery disease (2). In patients under- going coronary artery bypass grafting procedures, perioperative myocardial ischemia has been associ- ated with an increased risk of perioperative myocar- R, 'pheral vascular disease (1) and in those with Presented in part at the American zyxwvutsr Society of Anesthesiologists Annual Meeting, San Francisco, California, October 1988. Accepted for publication June 5, 1992. Address correspondenceto Dr. Ellis, Department of Anesthesia and Critical Care, The University of Chicago, 5841 South Maryland Avenue, Box 428, Chicago, IL 60637. dial infarction (3). In patients undergoing noncardiac surgery, ST-segment depression detected by Holter monitoring after surgery predicts a ninefold increase in the risk of ischemic postoperative events (4). Because myocardial ischemia is often not detected by clinicians when viewing oscilloscopes (5), automated vigdance of the ST segments of the electrocardiogram (ECG) has been proposed as a method to alert clini- cians to changes in the ECG (6,7). Several instruments that continuously monitor de- viation of the ST segment at a defined point in the cardiac cycle are now available. Although this tech- nology has been used to define demographic predic- 764 Anesth Analg 1992;75:764-72 01992 by the International Anesthesia Research Soaety 0003-2999/92/$5.00