Effect of hemodilution, blood loss, and transfusion on the predicted fibrinogen, F1.2, thrombin-antithrombin (TAT), and thrombin generation rates in subject 2. Each row shows the measured levels (filled circles) versus predicted levels (dotted line) for fibrinogen (first column), F1.2 and TAT (second column), and the resulting thrombin generation rate (third column). Model E assumes no hemodilution; blood loss or transfusion is occurring, resulting in a poor fittothe fibrinogen data and underestimation of the thrombin generation rate. Model F incorporates hemodilution but not blood loss or transfusion, resulting in a better fit with fibrinogen during early CPB and a higher estimate of thrombin generation. Model G incorporates hemodilution, blood loss, and transfusion in the circulatory model, resulting in the best fit between measured and predicted fibrinogen, F1.2 and TAT, and the most accurate estimate of in vivo thrombin generation. The solid black box at the top of the figure indicates the period of cardiopulmonary bypass (CPB).