TABLE 1

Comparison of predicted drug-drug interaction (DDI) liabilities from in vitro data to in vivo clinical studies

Hepatobiliary clearances, after in vitro to in vivo extrapolation (IVIVE), can be used to identify the rate-determining step (RDS) of a drug, such as if the CL_{met + bile}/CL^s_ef ratio is > or < than the tipping point (eq. 2), then the drug will have RDS_uptake or RDS_all, respectively. If a drug has RDS_uptake, then the PI_{met + bile} can be quantified (eq. 5) to predict when a significant DDI should be expected owing to inhibition of metabolic/biliary efflux clearance. An expanded analysis is shown in Supplementary Table 1.

Drug	Hepatobiliary Clearance (ml/min per kg)				fu_bCL^s_in/Qh^{^a}	CL_{met + bile}/CL^s_ef	Tipping Point	PI_{met + bile}	RDS In Vitro	RDS In Vivo	Reference
Drug	CL^s_in	CL^s_ef	CL_met	CL_bile	fu_bCL^s_in/Qh^{^a}	CL_{met + bile}/CL^s_ef	Tipping Point	PI_{met + bile}	RDS In Vitro	RDS In Vivo	Reference
Atorvastatin	61^{^b}	24.9	58.3	4.3	0.12	2.52	3.59	—	All	Uptake^{^c}	Varma et al. (2014)
	1194^{^d}	24.9	58.3	4.3	2.27	2.52	1.22	>51%	Uptake		Camenisch and Umehara (2012)
	405^{^e}	24.9	58.3	4.3	0.77	2.52	2.26	>10%	Uptake		Kunze et al. (2015)
	198^{^b}	359	64.6	11.8	0.80	0.21	2.22	—	All		Maeda et al. (2011)
	198^{^b}	57.7	64.6	11.8	0.80	1.32	2.22	—	All
Bosentan	132^{^b}	28.9	19.5	5.8	0.36	0.87	2.95	—	All	Uptake^{^c}	Varma et al. (2014)
	142^{^d}	28.9	19.5	5.8	0.38	0.87	2.90	—	All		Jones et al. (2012)
	1117^{^e}	28.9	19.5	5.8	3.02	0.87	0.99	—	All		Yoshikado et al. (2017)
	35^{^b}	12.1	—	39^{^f}	0.02	3.24	3.93	—	All
	2035^{^g}	14	—	5.0^{^f}	1.09	0.36	1.91	—	All
Repaglinide	166^{^b}	63.6	128	0.3	0.19	2.01	3.35	—	All	Uptake^{^c} all^{^h}	Varma et al. (2014)
	1983^{^d}	63.6	128	0.3	2.32	2.01	1.21	>40%	Uptake		Jones et al. (2012)
	1151^{^e}	63.6	128	0.3	1.35	2.01	1.71	>15%	Uptake		Yoshikado et al. (2017)
	299^{^b}	223	125	0.0	0.22	0.56	3.27	—	All
	3671^{^g}	352	125	0.0	2.73	0.35	1.07	—	All

↵^a The fu_b and Q_h values as noted in each reference were used for analysis; note that fu_b may vary for the same drug across different references.
↵^b In vitro quantified + IVIVE.
↵^c RDS_uptake was established in vivo for atorvastatin and bosentan since there was no significant AUC change to victim drugs when coadministered with intravenous itraconazole (CYP3A inhibitor) which resulted in 33% and 73% CYP3A inhibition, respectively (Maeda et al., 2011; Yoshikado et al., 2017). Midazolam, a CYP3A probe, was used to assess magnitude of CYP3A inhibition. RDS_uptake was established for repaglinide via whole-body physiologically based pharmacokinetic (PBPK) modeling of complex transporter- and enzyme-mediated DDIs (Varma et al., 2013).
↵^d In vitro quantified + IVIVE + empirical scaling factor for active uptake transport (individual scaling factor).
↵^e In vitro quantified + IVIVE + empirical scaling factor for active uptake transport (geometric mean scaling factor).
↵^f Composite CL_{met + bile}.
↵^g Fitted parameters from in vivo using PBPK model.
↵^h RDS_all was established in vivo for repaglinide even though there was no change to systemic AUC by intravenous itraconazole because CYP2C8 is the major hepatic drug-metabolizing enzyme (Yoshikado et al., 2017). In a different study, oral trimethoprim, a selective CYP2C8 inhibitor, increased repaglinide AUC by 1.8-fold (Kim et al., 2016).