Computerized vectorcardiography for on-line cardiac monitoring in non-cardiac surgery

Sammanfattning: Computerized Vectorcardiography for On-line Cardiac Monitoring in Non Cardiac Surgery Per Gannedahl Department of Surgical Sciences, Karolinska Institute and Karolinska Hospital, Stockholm, Sweden. Perioperative cardiac morbidity (PCM) remains high in risk groups such as peripheral vascular surgery. Preoperative assessment has not succeeded in identifying the individual patients most susceptible of PCM. Computerized vectorcardiography (VCG), a novel continuous on-line method for ischaemia monitoring was evaluated for perioperative use. Special emphasis was placed on its predictive capability for PCM in risk patients. Spatial changes in QRS (QRS-VD) and ST (STC-VM) as well as absolute ST changes (ST-VM) were analysed. The influence of anaesthesia, surgery and positional changes on the VCG recordings was investigated in healthy patients undergoing open cholecystectomy (n=9), laparoscopic cholecystectomy (n=12) and inguinal hernial repair (n=5). The ST variables remained almost unchanged throughout, while the QRS-VD was significantly increased in the left lateral and sitting positions, when changing from the Trendelenburg to the reversed Trendelenburg positions and during pneumoperitoneum for laparoscopic surgery. The scalar 12-lead ECG derived from the VCG (dECG) was compared with simultaneously recorded standard 12 lead ECGs (sECG) in patients after elective surgery of the abdominal aorta (n=17), and subjects with acute anterior (n=18) and inferior (n=18) myocardial infarctions. Qualitative analysis showed good agreement between the methods, while the quantitative comparisons revealed some minor differences, mainly related to amplitude. The predictive value of VCG for subsequent PCM was determined in patients undergoing elective abdominal aortic surgery (n=38). The VCG changes were recorded during surgery and for 48 hours postoperatively. The recordings were analysed retrospectively and related to the 13 patients suffering from PCM, as detected on clinical grounds. No patients without VCG signs of ischaemia suffered from PCM. Thirty patients had VCG signs of ischaemia, yielding a sensitivity of 100%, which was 50% more sensitive than a three-lead ECG and almost 100% better than a single V5-lead. However, the specificity of VCG was only 32%. Calculations of the duration of cumulated ischaemia revealed that a critical limit of 260 minutes of ischaemia during two consecutive hours could be established. Applying this limit, the prediction of subsequent PCM could be made with a sensitivity of 85 %, specificity of 80% and positive and negative predictive values of 69 and 91 % respectively. If used on-line this criteria would have identified 11/13 patients subsequently suffering from PCM to be at risk. The dECG (12 lead) also showed high predictive values, but dECG signs of ischaemia appeared later than for the VCG (mean 78 minutes, range 0-38 hours). The VCG signs of ischaemia were significantly increased in the patients with PCM already intraoperatively. The degree of intraoperative and postoperative ischaemic VCG signs was found to be significantly correlated. VCG is a useful method for continuous perioperative ischaemia monitoring. In patients with a high risk of PCM, VCG has the potential to predict subsequent development of cardiac events with a high degree of accuracy. Only in some specific positions and during pneumoperitoneum should sole QRS changes be interpreted with some caution. The dECG from the VCG is reliable for clinical use. Key words: Anesthesia; electrocardiography; intraoperative monitonng; laparoscopy; myocardial ischemia; vascularsurgery; vectorcardiography. Stockholm 1997 ISBN 91-628-2466-X