Applications of sensitivity analysis to a physiologically based pharmacokinetic model for carbon tetrachloride in...

Physiologically based pharmacokinetic (PBPK) models developed from gas uptake experiments have been used to estimate metabolic parameters for volatile organic compounds. Due to the potential application of PBPK models to estimate metabolic bioactivation constants in humans, it is important to understand the complex nature of these models and the resulting estimates. Adult male F344 rats (165-205 g) were individually exposed to carbon tetrachloride (CCl4) in gas uptake systems. Three rats at each concentration were exposed for 6 hr to initial concentrations of 25, 100, 250, and 1000 ppm CCl4. Partition coefficient determinations were performed by the vial equilibration technique and used as model inputs. Computer optimizations with the means of each initial chamber concentration at each time point resulted in an estimate of Vmax of 0.11 mg/hr (Vmaxc = 0.37 mg/hr/kg) and Km of 1.3 mg/liter. To determine the effect of individual animal variation in Vmax, optimizations were also performed with the mean +/- SD, resulting in Vmax estimates of 0.09 and 0.12 mg/hr, respectively. Similar analysis resulted in Km estimates of 0.98 and 1.58 mg/liter. The results of the sensitivity analysis were concentration dependent for CCl4. These results show Vmax and Km to be most accurately detected at lower initial chamber concentrations. Results of the sensitivity analysis at the lowest concentration established the following model input hierarchy: blood to air partition > fat partition and fat volume fraction > slowly perfused partition, ventilation rate, cardiac output, fat blood flow percentage > liver blood flow percentage and slowly perfused blood flow percentage. Further sensitivity analysis determined Vmax and Km to be highly correlated when using gas uptake technology and point to the need to an independent estimate for either constant. In summary, the application of sensitivity analysis to PBPK modeling resulted in an increased understanding of factors governing the estimation of metabolic parameters.