In this study, we compared continuous cardiac output (CO) obtained from the femoral arterial pressure by simulation of an aortic input impedance model [model-simulated cardiac output (MCO)] to thermodilution cardiac output (TDCO) determined by bolus injection during liver transplantation. Both variables were measured in 39 adult patients (13 females) every 10th minute during liver transplant surgery. Paired measurements were compared during the 4 phases of surgery-dissection, anhepatic phase, early reperfusion (the first 15 minutes after reperfusion), and late reperfusion (15-60 minutes after reperfusion)-without the detection of any significant difference between the 2 estimates of CO. TDCO ranged from 2.3 to 17.2 L/minute, and the bias (the mean difference between MCO and TDCO) prior to calibration was ฯช0.4 ฯฎ 1.6 L/minute (mean ฯฎ standard deviation; 1309 paired measurements; 95% limits of agreement: ฯช3.4 to 2.6 L/minute). After calibration of the first determined MCO by the simultaneously determined TDCO, the bias was 0.1 ฯฎ 1.5 L/minute, with 57% (n ฯญ 744) of the comparisons being less than 1 L/minute and 35% (n ฯญ 453) being less than 0.5 L/minute; this was independent of the level of CO, and the mutual correlation coefficient was 0.812 (P ฯฝ 0.001). This study indicates that during liver transplantation surgery, MCO reflects TDCO throughout the operation. Thus, for CO, this less invasive method appears to provide a reliable uninterrupted measurement during orthotopic liver transplantation.