UCSF

Complex drug-drug interactions are defined as those in which both metabolic enzymes and xenobiotic transporters are implicated as clinically significant determinants of drug disposition. Both metabolic enzymes and xenobiotic transporters have the potential to contribute to clearance pathways (i.e. metabolic, renal or biliary elimination) and bioavailability-related processes (i.e. drug absorption, intestinal metabolism, or first pass hepatic elimination). Transporters have the unique ability to influence the distribution of drug throughout the body, in addition to influencing intestinal drug absorption or drug clearance via renal or biliary routes. Thus, characterization of the contributions of metabolic enzymes and xenobiotic transporters is crucial in anticipating any potential alterations in drug exposure due to a drug-drug interaction, pharmacogenomic or disease state variance of activity or expression of relevant metabolic enzymes or transporters.

Predictions of drug-drug interactions are routinely conducted based on results of in vitro metabolic enzyme or xenobiotic transporter inhibition studies. However, translating such results to clinical significance continues to challenge the field, particularly for transporter-mediated interactions since the susceptibility of a drug to transporters in vitro does not always translate to clinically significant in vivo involvement and due to a lack of specific and clinically validated index substrates, inhibitors and inducers for major xenobiotic transporters. The objective of this research was to provide a framework for recognizing transporter involvement in clinical drug-drug interactions, grounded in basic pharmacokinetic theory.

Since xenobiotic transporters can allow (or disallow) substrates access to various tissues throughout the body, it was recognized that significant xenobiotic transporter interactions are accompanied by changes in volume of distribution, in addition to potential changes in clearance, which can result in counterintuitive changes in mean residence time and terminal half-life. Metabolic interactions are not expected to result in any volume of distribution changes and this hypothesis was extensively evaluated via examination of 72 intravenous metabolic drug-drug interaction studies with clinically recommended index substrates and inhibitors. The results indicate that volume of distribution is almost always unchanged in strictly metabolic interactions with marked exposure changes, resulting in changes in mean residence time and half-life that are equal but opposite to clearance changes, further highlighting that volume and clearance are indeed independent parameters.

Understanding that metabolic interactions do not result in volume of distribution changes can allow for estimation of bioavailability changes in oral drug-drug interactions, where the extent of change in apparent volume of distribution will reflect changes in bioavailability alone due to unchanged volume of distribution. Such estimates of changes in bioavailability can subsequently be utilized to differentiate changes in clearance alone from measures of apparent clearance following oral dosing. This approach can also be utilized to predict if an overall exposure change for oral drug-drug interactions is primarily due to changes in systemic clearance versus bioavailability.

To identify clinically significant intestinal transporter interactions, it was demonstrated that alteration of intestinal transporter activity or expression will result in significant changes in absorption rate, and such changes should always be used to implicate transporter involvement in vivo. Inhibition of apical efflux transporters result in decreased absorption time, as efflux transporter-mediated drug cycling between the enterocyte and gut lumen is prevented, while efflux transporter induction results in prolonged absorption time, as reflected in values of mean absorption time and time to maximum concentration.

Analyses of clinical data, such as examining changes in volume of distribution following intravenous dosing, changes in absorption rate following oral dosing, and examining the relationship between clearance changes and half-life and mean residence time changes, can confirm transporter involvement of purported complex drug-drug interactions. Such an approach was utilized to critically evaluate the purported clinical significance of the efflux transporters P-glycoprotein (P-gp) and Breast Cancer Resistance Protein (BCRP) in the disposition of apixaban, as has been indicated throughout the literature and even on the apixaban FDA label. Rational examination of all published apixaban clinical drug-drug interaction studies, using the proposed basic clinical pharmacokinetic methodologies, does not support the clinical significance of the efflux transporters P-gp nor BCRP in apixaban disposition. In fact, inhibition or induction of intestinal metabolism via cytochrome P450 3A4 (CYP3A4) can account for all exposure changes of clinically significant drug-drug interactions, and lack of intestinal CYP3A4 inhibition can explain all studies with no exposure changes.

Understanding the utility and limitations of experimental systems, as well as the inherent assumptions of the pharmacokinetic equations utilized to translate such results, is crucial in translating in vitro or in situ experimental information to an in vivo prediction of drug disposition. For instance, there is limited benefit to using measurements of unbound liver-to-blood partitioning (Kpuu) to improve predictions of drug-drug interactions, as DDIs can adequately be predicted by the Extended Clearance Model without any measurements of intracellular drug concentrations, a difficult task hindered by experimental variability. Further, the recognition that Kpuu has inherently assumed the well-stirred model implies that such approaches cannot account for the nuances of intracellular drug distribution. Finally, recognition that clearance calculations based on extraction ratio have inherently assumed the well-stirred model further highlights the importance of understanding the assumptions inherent in basic pharmacokinetic relationships that are universally utilized to characterize clinical drug disposition.