Nonlinear Regression Case Study Pharmacokinetic Modeling of a New Chemical Entity

NONLINEAR REGRESSION CASE STUDY PHARMACOKINETIC MODELING OF A NEW CHEMICAL ENTITY 

The parameter estimates for the 2-compartment model are presented in Table 5.14 with the observed and model fitted data plotted in Fig. 5.16. A plot of observed versus predicted concentrations showed that at higher concentrations there was less agreement between the observed and predicted concentrations. Although the parameter estimates were estimated with good precision, a clear trend in the residual plot was observed. The model predictions tended to overpredict the 24 and 48-h concentration data and there was a slight discrepancy at 72 h although these were only a few data points. The residual plot demonstrated the inadequacy of the 2-compartment model. 

another compartment was added to the 2-compartment model creating a 3-compartment model. The starting value for the new intercompartmental clearance term was the intercompartmental clearance under the 2-compartment model, while the volume of the new compartment was set equal to 1/10, the volume of Compartment 2. The final parameter estimates are shown in Table 5.15 while Fig. 5.17 shows the model fit overlay plot. 

The addition of a 3-compartment model dropped the Akaike Information Criterion (AIC) from 5.788 to 5.134. Although the model fit under the 3-compartment model was better than the 2-compartment model, a trend in the weighted residuals versus time plot was still apparent indicating that the model tended to underpredict (from 1 to 12 h), then overpredict (from 20 to 

Parameter Model term Units Value Error CV (%) 

---