Impedance cardiography (IC) is a noninvasive, simple to use method of cardiac output (CO) determination. A prospective evaluation of IC monitoring was performed in 50 patients undergoing noncardiac surgery. IC CO measurements (NC-COM3-Revision 7, BoMed Manufacturing) were compared to simultaneous measurements of thermodilution (TD) CO to assess the validity of this technique for intraoperative cardiac monitoring. Adequate impedance signals could not be obtained in 7 of the 50 patients. IC CO measurements were highly correlated to TD CO (P < .005), with a correlation coefficient r = 0.84. Bias analysis, however, indicated clinically significant disagreement between the two techniques. IC CO tended to underestimate TD CO (mean bias = -0.41 L/min) and the SD of the bias was 1.0 L/min (95% level of agreement 1.6 to -2.4 L/min). Trending data showed IC to accurately track the direction of TD CO changes but to underestimate their magnitude (r = 0.60, intercept -0.7 L/min, slope 0.47). Factors that may have impaired the performance of IC in this study include the high prevalence of cardiac disease in the study population and electrical noise in the operative setting. Further development of IC appears warranted if it is to prove useful as an intraoperative cardiac monitor.