Hepatic clearance models: comparison of the dispersion and Goresky models in outflow profiles from multiple indicator dilution rat liver studies.

The multiple indicator dilution (MID) technique is often used for investigation of the kinetic behavior of substrates and metabolites in eliminating organs. The present study was a systematic comparison of the utility of the Goresky model (GM) (a structural model) and the mixed-boundary dispersion model (DM) (a stochastic model) in the interpretation of influx, efflux, and removal (sequestration) coefficients, with data generated from rat liver-perfusion/MID studies. Although the GM and the DM are equivalent in their descriptions of membrane transport, they differ in their classifications of the dispersion of blood-borne elements. For the DM, the dispersion is an inverse Gaussian distribution of vascular transit times; for the GM, it is accounted for by the dispersion observed among noneliminated reference indicators (e.g. labeled red blood cells, albumin, sucrose, and H2O) or the derived reference. In this study, previously published rat liver-perfusion/MID data obtained for the glutathione conjugate of bromosulfophthalein and hippuric acid, compounds that exhibit saturable carrier-mediated transport, with the GM were reanalyzed with the two-compartment DM. When the fitted values for volume and transfer coefficients were compared, good correlation was found between the fitted vascular volume for the DM and the vascular volume for the reference indicator for the GM. The influx coefficients were generally similar between the models, but improved correspondence was observed when the DM was modified to include the large-vessel transit time. In contrast, the efflux and sequestration coefficients obtained for the DM did not correspond well to those from the GM. The disagreement was due, in part, to differences in the interpretation of the late-in-time component of the reference transit time distribution curve, which was not described well by the DM. Consequently, the residence time distribution and the relative dispersion were underestimated by the DM.