Figure 3. Observed and predicted [Greek small letter epsilon]-aminocaproic acid (EACA) concentrations for the patients in groups I and II using a model based on data from groups I, II, and III either without (A) or with (B) corrections for the effects of cardiopulmonary bypass (CPB), as described in the text and Table 4. Observed and predicted EACA concentrations for group III are provided using the model based on data from groups I, II, and II without CPB corrections (C) and using the same data with CPB corrections (D). Data are plotted on a semilog scale. Dashed line = time of CPB. CPB corrections had little effect on model predictions for groups I and II, in which EACA was administered after CPB. Conversely, the CPB corrections significantly improved model predictions in group III, in which EACA was administered before, during, and after CPB. Time along the abscissa is measured from administration of the loading dose in all cases. Note that observed and predicted value curves are "flatter" in D than in C, confirming a more consistent fit. A simple explanation for the seemingly parallel line is variability in EACA volumes of distribution.