BDDCS, the Rule of 5 and Drugability (2024)

The publisher's final edited version of this article is available at Adv Drug Deliv Rev

2.1. Historical Development of BDDCS

The BDDCS was an outgrowth of the Biopharmaceutics Classification System (BCS), which developed from the seminal 1995 paper of Amidon et al. [7] that led to the FDA BCS Guidance in 2000 [8]. Wu and Benet recognized that the overwhelming majority of BCS Class 1 and Class 2 drugs were eliminated in man primarily via metabolic processes, while the overwhelming majority (41 of 42) of BCS Class 3 and Class 4 drugs classified at that time were primarily eliminated in man unchanged in the urine and bile [1]. Recognizing that the BCS classification may be predictive of drug disposition, in addition to drug absorption, Wu and Benet evaluated enzymatic and transport characteristics of drug substrates with respect to the BCS classification. Our laboratory had published many papers related to metabolism and transporters and was the first to propose the potential and relevance of transporter-enzyme interplay [9]. We incorporated the many studies from our laboratory evaluating enzymes and transporters from molecular, in vitro animal, in vivo animal, and human studies leading to the BDDCS proposal [1].

It is important to recognize that the 22 predictions made by Wu and Benet with respect to the BDDCS system were based on observations, not theory, primarily from studies investigating the relevance of transporters, enzymes and transporter-enzyme interplay, particularly for immunosuppressives [10–15]. These studies served as the basis for the prediction of the relevance of transporters for the various BDDCS classes as depicted in Fig. 1 for orally dosed drugs [16].

Open in a separate window

Figure 1

BDDCS Classes for Orally Dosed Drugs, Percentage of Drugs in Each Class and Prediction of Transporter Effects [16]

For the 153 drugs initially classified in the BDDCS system by Wu and Benet [1], we were unable to identify any clinically relevant transporter effects in the gut or the liver for the BDDCS Class 1 drugs. This prediction from 2005 was not based on theoretical concepts, but rather findings that were familiar to us as biopharmaceutical scientists with knowledge of drug metabolism, transporter effects and pharmacodynamics. In that paper [1], Wu and Benet cautioned that “there will always be exceptions to the broad, general rules presented here”. Yet even when the BDDCS classification was expanded to more than 900 drugs [17] and most recently an additional 175 drugs were added [18], the predictions in Fig. 1 for the BDDCS Class 1 drugs holds remarkably well and we do not know of any class 1 drugs that require a dosage change as a result of transporter inhibition or induction. Recently, Varma et al. [19] have suggested that two statins, fluvastatin and cerivastatin, classified as BDDCS Class 1, do exhibit rate limited uptake into hepatocytes as a function of OATPs. But, their suggestion is not supported, and is in fact contradicted, by clinical data. Niemi and co-workers [20,21] report that OATP1B1 polymorphisms that have been shown to affect the pharmaco*kinetics of all of the BDDCS Classes 2, 3 and 4 statins, do not affect the pharmaco*kinetics of the BDDCS Class 1 statin, fluvastatin. Cerivastatin was removed from the market before any such evaluation was carried out. Varma et al. [19] have fallen into the trap noted in Fig. 1; BDDCS Class 1 compounds, which represent 37.5% of classified drugs [17, 18], can be shown to be substrates of transporters, but these transporter effects are clinically insignificant. The disposition of the remaining 62.5% of classified drugs may be modified by transporters.

Similarly, the predictions for transporter effects for the highly metabolized Class 2 drugs, as depicted in Fig. 1, were based on observations and studies in our laboratory. The prediction that efflux transporter effects predominate in the gut was primarily based on the work of Dr. Carolyn Cummins [12, 15, 22–25]. These studies showed that when an extensively metabolized drug is a substrate for an efflux transporter, absorption of parent drug as well as the extent of intestinal metabolism are modified by efflux transporters. In particular, inhibition of efflux transporters has been shown to decrease the extent of intestinal metabolism, since efflux inhibition prevents a recycling mechanism that allows the drug multiple times to access a metabolizing enzyme. However, we were unable to identify any uptake transporter effects in the intestine for BDDCS Class 2 drugs. The recognition of the potential importance of both uptake and efflux transporters in the liver was based on the previously referenced studies of Drs. Chi-Yuan Wu and Yvonne Lau [13–15]. The prediction that BDDCS Class 3 and Class 4 drugs would require an uptake transporter in the intestine was based on the recognition that these poorly metabolized, poorly permeable drugs would require uptake transport to achieve a relatively rapid rate of absorption, but then once absorbed, these drugs could be substrates for efflux transporters in the intestine as well as the liver.

In general, now with the BDDCS Classification expanded to more than 1,100 drugs and active metabolites (primarily active compounds from prodrugs) [17,18], the predictions in Fig. 1 and the relationship between intestinal permeability and metabolism hold up quite well, such that we do not know of any drugs predicted to not have a clinically relevant transporter effect (37.5% of drugs) that exhibit a relevant transporter effect in the clinic, thus leading to the potential drug-drug interaction (DDI) predictability of BDDCS. That is, DDIs will only be metabolic in the intestine and the liver for the BDDCS Class 1 drugs. For the Class 2 drugs, DDIs can include metabolic, efflux transporter and efflux transporter-enzyme interplay in the intestine, as well as metabolic, uptake transporter, efflux transporter and transporter-enzyme interplay in the liver. The primary DDIs for BDDCS Class 3 and 4 drugs will be uptake transporters, efflux transporters and uptake-efflux transporter interplay. However, we always caution that a simple four category system won’t predict every interaction. BDDCS doesn’t propose that every drug in a class will be substrates or not substrates for uptake and efflux transporters, rather, BDDCS serves to prioritize which interactions should be investigated, and in what order.

2.2 The Differences between BDDCS and BCS

“The major differences between BDDCS and BCS relate to their purpose and the measurement for classification, as previously presented in Table 1 [26]. The purpose of BCS is to characterize drugs for which products of those drugs may be eligible for a biowaiver of in vivo bioequivalence studies. The purpose of BDDCS is to predict drug disposition and potential drug-drug interactions in the intestine and the liver and potentially the kidney and brain. Both BCS and BDDCS use solubility as one of the two classification criteria. The solubility parameter utilized, referred to as ‘US FDA solubility’, estimates the ability of a drug at its highest dose strength to completely dissolve in 250mL of water over a pH range of between 1 and 6.8 at 37°C” [8]. (Note: with the universal use of plagiarism detection software by journals today, there is a great deal of unjustified concern about self-plagiarism even when the text indicates that the sources are earlier publications of the authors. Thus, rather than reword the same text to which we are referring, we have just included many quotation marks even though the wording may not be an exact quote due to the inclusion of attributions and references.) The volume parameter (250 mL) estimates the volume of a glass of water, ideally reflecting the minimum volume available for dissolution of a drug product in the stomach and intestine, while the pH range reflects the pH range a drug may encounter in the stomach and intestine prior to absorption. “For a drug to be considered highly soluble in both classification systems, the drug from its highest strength regulatory approved dosage form must go completely into solution at its lowest solubility over this pH range in 250mL of water. As we have recently noted, US FDA solubility is a property of the drug in a formulation and is not an intrinsic property of the active pharmaceutical ingredient itself” [27]. There is no inherent accuracy to the 250ml value chosen for the FDA criterion. However, it seems to usefully differentiate Class 1 and Class 2 drugs in terms of their transporter susceptibility and thus we continued the BCS solubility criterion into BDDCS. “The second classification parameter and where the two systems differ, is related to intestinal permeability. In BDDCS, predictions are based on intestinal permeability rate, which was found to be related to the extent of drug metabolism. In BCS, biowaivers are based on the extent of intestinal absorption, which in a number of cases does not correlate with intestinal permeability rate” [8].

2.3 The Scientific Basis for BDDCS

As noted above, the BDDCS classification and the predictions from the 2005 paper [1] were based on observations and knowledge of pharmaco*kinetics and pharmacodynamics for drugs on the market. The recognition of the correlation between intestinal permeability rate and the extent of metabolism preceded an explanation for these findings. One might ask, Why should jejunal intestinal permeability rate in humans, the measurement from the initial human BCS studies, predict the extent of metabolism? and, Why should a kinetic measure, permeability rate, predict a thermodynamic outcome, extent of metabolism? We hypothesize now that high permeability rate compounds are readily reabsorbed through the kidney lumen and from the bile facilitating multiple accesses to the metabolic enzymes. This would be particularly important for metabolically eliminated drugs with a low hepatic clearance (e.g., diazepam and letrozole). The explanation for letrozole, previously presented [26], is repeated here, “For example, consider the BCS/BDDCS Class 1 drug letrozole. This completely oral available drug is primarily eliminated by CYP3A4 and CYP2A6 enzymatic processes with less than 4% of the dose excreted unchanged in the urine. Letrozole is only 60% bound to plasma proteins and thus it might be expected, based on glomerular filtration rate and fraction unbound, that renal clearance should approach 48mL/min. Yet, the total clearance of letrozole is only 45mL/min with less than 4% excreted unchanged. Thus, this high-permeability rate compound is reabsorbed in the kidney tubules (and possibly from the bile) with the major route of elimination being metabolic processes. The rationale for the correlation between intestinal permeability rate and the extent of metabolism appears to be based on the fact that high permeability rate compounds are reabsorbed from potential unchanged drug excretion routes in the body and thus can only be eliminated through metabolism.”

One may also ask, “Why should solubility play such a prominent role in differentiating the effect of transporters between BDDCS Class 1 and Class 2 drugs?” As noted above, both BCS and BDDCS use solubility as one of the two classification criteria and that US FDA solubility is a property of the drug in a formulation and is not an intrinsic property of the actual pharmaceutical ingredient itself [27]. Similarly, Varma et al. [19] note that “Solubility is a fundamental principle for oral absorption as only drug in solution has the ability to permeate across enterocytes…however, it is not directly related to drug clearance.” We emphasize again that the predictions in Fig. 1 concerning the importance of transporters for BDDCS Class 2 drugs, versus BDDCS Class 1 drugs, were based on observation, not theory. As an explanation, we very recently noted [28] “For Class 1 compounds, which are highly soluble, highly permeable and extensively metabolized, the passive permeability at concentrations unrestricted by solubility appears to overwhelm any potential transporter effects. In this situation, Class 1 compounds may be substrates of transporters in cellular systems, but transporter effects will be clinically insignificant in the gut, liver or brain. For example verapamil was shown to be a substrate of Pglycoprotein (P-gp) in the MDCK-MDR1 cellular system, but it exhibits no clinically significant P-gp transporter effects in the intestine, liver and brain.” It has been suggested that given its high permeability, verapamil overloads the ATP-dependent transporter capacity of P-gp, effectively acting as a P-gp inhibitor [29]. Our latest thinking is that solubility is a characteristic of a drug substance that subsumes a number of individual characteristics that we and others have not yet been able to identify or quantify that are determinants of drug disposition. Our latest analyses suggest that a 100mg (or very slightly poorer, 50mg) in 250ml water over the pH range 1–6.8 adequately predicts BDDCS class, independent of highest approved dose strength. And that this pH range is important, so we would not reclassify acids that only fail the solubility criterion at pH 1, or suggest that a drug may be a different BDDCS class at a lower dosage. We emphasize again our belief that there is nothing magical about the 250ml volume and that for an NME, of course, it could be prohibitive to test 100mg. It is the ratio that appears defining. Thus, high solubility compounds would exhibit full solubility over the pH range 1–6.8 of ≥ 0.4mg/ml, while poor solubility compounds would exhibit solubilities < 0.4mg/ml.

3.1 Marketed Drugs’ Characteristics

“To facilitate use of the BDDCS system for making predictions for marketed drugs, in 2011 we compiled the BDDCS classification for 927 drugs, which included 30 active metabolites, primarily the active species from prodrugs” [17]. More recently, Hosey et al. [18] incorporated an additional 175 drugs into the system and amended the classification of 11 drugs from the previous compilation. Our analysis here evaluates the more than 1100 drugs compilation as amended, where we have excluded 14 Class 0 drugs (those drugs where the extent of metabolism and BDDCS class would change dependent upon urine pH), 31 active metabolites (when the active metabolite is not also marketed in a regulatory approved dosage form) and 11 drugs for which we could not confirm an approval history. Thus the Ro5-BDDCS analysis here was carried out on 1064 parent drugs, of which 852 drugs are administered orally. In our initial compilation [17] “where the lowest measured solubility was found in the literature, this value was reported. Measured values are also reported for the percentage excreted unchanged in the urine, LogP and LogD 7.4, when available” [17]. For the more than 1100 compounds in both compilations [17,18] we “determined the in silico parameters for predicted Log solubility in water (both dependent and independent of pH), calculated LogP, polar surface area, and the number of hydrogen bond acceptors and hydrogen bond donors for the active moiety, thereby allowing comparison between in silico and experimentally measured values. When comparing the in silico parameters across the four classes, the most distinct difference was noted between Class 2 and Class 3 compounds [17]. However, CLogP and in silico solubility parameters for the Class 1 drugs were found to have intermediate values between those for Class 2 and Class 3 drugs, and not very different from the values for Class 4 drugs. We emphasized this failure of the in silico parameters to efficiently predict whether a drug could be Class 1, believed by many to be the most desirable because of high solubility and high permeability, versus Class 4 drugs that are low solubility and low permeability.”

3.2. Analysis of BDDCS Compounds Using Ro5

Since the 2011 compilation [17] included the four Ro5 criteria for each of the parent drugs, and it was easy for us to determine the Ro5 criteria for the additional compilation [18], it is possible to evaluate these marketed drugs in terms of the Lipinski criteria. In Table 2 we list the results for the 852 drugs administered orally to achieve systemic concentrations and the 212 drugs not administered orally to achieve systemic concentrations (N = 1064 compounds as detailed above). As noted in Table 2, most compounds that exhibit 2 or more Ro5 violations are natural products or natural product derivatives (certainly true for non-oral drugs). Strictly speaking, Ro5 criteria are not applicable to those compounds, but they were included in this evaluation for the sake of completeness. Ro5 violations were determined for molecular weight > 500, CLogP >5.0, sum of nitrogen and oxygen (N, O) atoms > 10 and hydrogen bond donors > 5. For the 322 orally administered Class 1 compounds, 35 exhibited a single Ro5 violation, five (bromocriptine mesylate, dabigitran etexilate, fosinapril, olmesartan medoxomil [all four essentially prodrugs] and reserpine) exhibited two Ro5 violations, and four drugs (acarbose, cyanocobalamin, everolimus and ivermectin) exhibited three Ro5 violations. As noted above, Lipinski’s Ro5 was developed to set “drugability” guidelines from an oral bioavailability perspective [6] for small molecules [3] and suggested that when a compound exhibited two or more Ro5 violations the compound would exhibit poor solubility and/or poor permeability. Thus, 9 of the 322 approved BDDCS Class 1 orally administered drugs (2.8%) would have been predicted to exhibit poor drugability. As noted above, when the Ro5 was developed, Lipinski specifically stated that the Ro5 only holds for compounds that are not substrates for active transporters [2,3] and little information was available concerning drug transporters at that time. As noted in Fig. 1, transporter effects are minimal in the gut and liver and clinically insignificant for BDDCS Class 1 compounds. Thus, it is not surprising, that the Ro5 performs so well for class 1 compounds, at least for drugs on the market. Furthermore, 6 of the 9 Ro5 violation drugs, bromocriptine (an ergot alkaloid derivative), reserpine (a rauwolfia alkaloid), acarbose and everolimus (both microbial products), ivermectin (derived from a fungi product) and cyanocobalamin (vitamin B12) are natural products or derivatives, and are specifically excluded from the Ro5 criteria. Only the three remaining prodrugs are chemically synthesized small molecule oral BDDCS Class 1 violators.

However, even for Class 2 through 4 BDDCS compounds, the Ro5 performs quite well for orally administered drugs, with only a slightly better predictability for the highly soluble Class 3 drugs versus the poorly soluble Class 2 and Class 4 BDDCS drugs. Thus, even when transporters may mediate the absorption of compounds (for some class 2 drugs and presumably all class 3 and 4 drugs), the Rule of 5 can predict oral absorption. Just because a drug is a substrate for a transporter does not mean that it will fall outside of the Rule of 5, rather, a drug that falls outside of the Rule of 5 and is orally available is assumed to be a substrate for a transporter. Highly soluble, highly permeable predicted to be metabolized Class 1 drugs appear to require adherence to Rule of 5 to be absorbed, while classes 2–4 may “break” the Rule of 5 and still be absorbed. However, orally administered compounds, even those that are substrates for transporters, rarely break Rule of 5. This is in marked contrast to Ro5 violations of two or more for all of the non-oral drugs, independent of BDDCS class, as noted in Table 2. This is logical, since many of these drugs are formulated for non-oral delivery because of difficulties of membrane permeability and solubility.

3.3. The Rule of Five: Cross- and Temporal Validation

Subtracting the 73 non-oral drugs exhibiting 2 or 3 Ro5 violations from the total 212 in Table 2, suggests that 139 (65.6%) of the non-oral drugs should have been orally available according to the Ro5. Of the 139, 118 compounds are high solubility (i.e., BDDCS class 1 or class 3), which suggests the possibility that some of these compounds could have been orally formulated if the sponsor had chosen to do so. An additional 67 (7.9%) of the 852 orally formulated drugs (Table 2) violate 2 or 3 of the Ro5 criteria; however, since most of these are natural products or natural product derivatives, Ro5 criteria would have not been applicable. This “cross-validation” analysis, far from diminishing the importance of the Rule of Five, substantiates the observation that, in drug discovery, no rule is absolute.

To temporally validate the Rule of Five, we evaluated Ro5 parameters for orally dosed drugs approved before and after (661 and 191, respectively) the first Ro5 paper was published (December 1997) as presented in Table 3. For the 191 orally dosed drugs approved after 1997 almost three times (2.8 fold) as many violated 2 or more Ro5 parameters as for the 661 orally dosed drugs approved before the Ro5 publications (for all classes 5.6% before and 15.7% after). This more than doubling was found for both the Class 1 and Class 2 drugs, although as previously noted in Table 2, the Class 1 drugs, those for which clinically relevant transporter effects are not expected, exhibited the fewest violations for metabolized compounds. In contrast to the 2.8 fold differentiation seen with orally dosed drugs, approved non-orally dosed drugs in all classes exhibited only a slightly greater percentage of Ro5 violations (all classes 32.4% before and 43.6% after). We add the caution previously raised that we only have data available for drugs that successfully received regulatory approval, and not the full range of new molecular entities examined by the industry.

3.4. Is There A Need to Update Ro5 Criteria?

The results for the two or more Ro5 violations for the 852 orally dosed drugs as presented in Table 2 suggest that the originally proposed criteria function quite well, especially for the BDDCS Class 1 drugs where little clinical relevance for transporters is observed. A few outlier categorizations can be seen. The first, for large molecular weight cyclic peptide immunosuppressants (i.e., cyclosporine, sirolimus, tacrolimus and everolimus) has been well recognized. Macrolide antibiotics (e.g. clarithromycin, erythromycin and roxithromycin) are also observed to show higher violation numbers. Once again, we note that the cyclic peptide immunosuppressants and the macrolide antibiotics are (for the most part) natural products and, as such, are not subject to Ro5 criteria. There is also a strong tendency for HIV protease inhibitors (e.g., atazanavir, lopinavir, nelfinavir and tipranavir) and tyrosine kinase inhibitors (e.g. lapatinib and nilotinib) to show this variance. Very recently, Leeson reported in this journal [30] that 14 of 19 (74%) HIV protease inhibitors and HCV drugs exhibited 2 or more Ro5 violations. We concur with the extent reported by Leeson. Among 12 HIV protease inhibitors we find that 7 exhibit two Ro5 violations, one (atazanavir) exhibit three violations and the remaining four exhibit one Ro5 violation. We include 9 HCV drugs in our listing, 8 of which exhibit 2 or 3 violations. Thus, 16 of the 21 (76%) HIV protease inhibitors and HCV drugs exhibit two or more Ro5 violations. This prevalence of HIV protease inhibitors and HCV drugs in the listing of drugs approved after the Ro5 publications reflects in part the increased percentages noted in Table 3.

It is obvious from the listings of the 661 oral drugs that came to the market prior to the publication of the Lipinski Ro5 papers [2, 3] that the criteria proposed did an outstanding job of characterizing oral drugability, especially considering that a large fraction of the Ro5 violators are natural products or natural product derivatives. This is true in spite of the fact that other definitions of the Ro5 criteria may be employed. For example, with respect to number of H-bond acceptors, is the ester oxygen an acceptor in C(=O)OR? What about the nitrogen in an amide, is that an acceptor? In essence, no to both questions, but because there is an occasional X-ray where this can be observed, there could be some disagreement. Furthermore, using the actual publication date of the Ro5 papers as the before/after divisor, does not consider that many approved drugs after the Ro5 publications could not have been influenced by Ro5. We only use this arbitrary time point to note that significantly more Ro5 violations are observed after the publication. In other words, the Ro5 papers did an excellent job of subsuming the characteristics inherent in good drugability at the time they were published. Subsequently, less emphasis on these Ro5 criteria is observed, yet still showing good drugability characteristics. We have already noted above the various classes of drugs exhibiting marked Ro5 relations and also note that the increased prevalence of prodrugs has been incorporated in drug discovery and development to overcome these Ro5 criteria.

There appears to be no differentiation in metabolic pathways and clearance patterns when small molecules enter into the beyond Ro5 chemical space. In addition, first pass gut metabolism appears to be of equal importance for oral drugs in all of the categories listed in Table 2, where metabolism is significant. Thus, we believe Ro5 criteria function as well today in defining drugability from an ADMET criterion as they did when first proposed by Lipinski and coworkers. As stated above, there is no competition between Ro5 and BDDCS; the purpose of BDDCS is to define and predict the disposition characteristics of NMEs whether they meet or violate Ro5 criteria.

4.1 Metabolism versus Excretion of Unchanged Drug in the Urine and Bile as the Major Elimination Route for an NME in Humans

The major, but simple, discovery from the BDDCS was the recognition that the jejunal intestinal permeability rate could differentiate metabolism versus excretion of unchanged drug as the primary route of elimination of an NME in humans [1]. It was then shown that in vitro permeability rate measures of an NME in cellular systems and even in non-biological membranes such as PAMPA would allow this prediction to be made before the NME had ever been dosed to animals or humans. Hosey and Benet [31] showed that depending on the system studied, the best differentiation between metabolized (BDDCS class 1 and 2) versus biliary and renal eliminated drugs (BDDCS class 3 and 4) would be obtained using labetalol as the reference compound when in vitro permeability rates were measured in the Caco-2 system, zidovudine when in vitro permeability rates were measured in the MDCK cell line and theophylline when the permeability rates were determined in the PAMPA cell line.

The performance measures for these reference compounds in the various permeability systems are presented in Table 4, taken from Hosey and Benet [31]. Sensitivity is a measure of the fraction of the drugs eliminated primarily by metabolism (BDDCS class 1 or 2) that are correctly predicted in the three systems. Specificity is a measure of the fraction of the drugs eliminated unchanged in the urine and bile (BDDCS class 3 or 4) that are correctly predicted by the three systems. Positive predictive value is a measure of the fraction of drugs that the three systems predict to be eliminated primarily by metabolism (by a high in vitro permeability rate) that are, in fact, eliminated by this mechanism while negative predictive value is a measure of the fraction of the drugs predicted to be eliminated primarily by biliary and renal excretion (by a low in vitro permeability rate) that are, in fact, eliminated by these processes.

However, measures of permeability in cellular and artificial membrane systems require experimental studies and it would be beneficial if such predictions could be based on in silico calculations. Table 5 from Hosey and Benet [31] provides a comparison between in vitro models and in silico predictions. Column 2 compares the ability to predict extensive metabolism versus elimination of unchanged drug in the three model systems, all exhibiting receiver operating curve areas greater than 0.9 versus evaluation using three in silico permeability rate models: ADMET Predictor MDCK [32], ADMET Predictor P_eff [32] and VolSurf+ Caco-2 [33]. In our opinion, at present these three in silico predictors of elimination by metabolism versus elimination as unchanged drug are not sufficiently accurate.

If we can compare the success of the Ro5 in silico parameters to predict drugability of approved orally administered drugs in Table 2 with the receiver operating curve measures in column 2 of Table 5 to differentiate metabolism versus non metabolism as the major route of elimination, one could suggest that the in silico drugability predictability is better. However, as we noted above, we only have the ability to evaluate drugability success for drugs that have been approved, i.e., by definition successful. As early as 2000, we showed that over 82% of the non-drug (chemically filtered) subset of molecules from the Available Chemicals Directory (over 123,000 molecules) were Ro5 compliant [6]. As we wrote, “The ‘rule of 5’ test cannot be used to discriminate between ‘drugs’ and ‘nondrugs’. This inability is due to the lack of significant differences in the distribution of the four properties examined in this test, i.e. MW, CLOGP, H-bond donors and H-bond acceptors.” [6] Thus, Ro5 is not an intrinsic metric for drugability, and it is not recommended as a tool to “enrich” a set of chemicals with drug-like entities from a random set of molecules, except with respect to oral bioavailability potential.

4.2 Extension of BDDCS to Predict Specific Transporter Effects and Quantitative Clearance Estimates

Ro5 criteria are designed to estimate the upper bounds in defining the pharmaco*kinetic space for oral dosing. BDDCS expands upon the Ro5 criteria to qualitatively predict route of elimination, the relevance of transporters and enzymes and potential DDIs. However, neither the Ro5 nor BDDCS provide quantitative predictions.

Recently, Camenisch and coworkers [34–36] have proposed an Extended Clearance Concept Classification System (ECCCS) to identify the rate-determinant hepatic clearance step for an NME and then provide a quantitative prediction of in vivo hepatic clearance. Varma et al. [19] also proposed an Extended Clearance Classification System (ECCS) to predict clearance mechanisms early in drug discovery. Furthermore, Sugiyama and coworkers propose an in silico classification method designated as CPathPred to predict the major clearance pathways of drugs [37, 38]. All of the above methods provide valuable insight into predictions of hepatic clearance both in terms of pathways and some for quantitative estimations. In contrast, BDDCS is not limited to predicting hepatic clearance only (this is true for Ro5 also), but a number of the advances noted in these further classification systems [19, 34–38], can be incorporated into BDDCS. We note, however, that Ro5 was not intended, nor used, for predicting drug clearance.

All three of the groups referenced above recognize that high molecular weight (>400) acids and zwitterions exhibit rate limited hepatic clearance as a function of OATP uptake into the liver, that is specifically designated as a separate class by Varma et al. [19]. This appears to be true for high permeability rate compounds that are metabolized such as atorvastatin and glyburide (BDDCS Class 2 compounds) and low permeability compounds that are excreted unchanged in the bile like valsartan and cefoperazone (BDDCS Class 4 and 3 compounds, respectively), but as we noted earlier this appears not to be true for BDDCS Class 1 compounds such as fluvastatin.

Thus, we would add to our BDDCS recommendations that all Class 2, 3 and 4 compounds with molecular weights >400 be tested for hepatic uptake by OATPs as being the rate limiting step for elimination of these compounds whether they are primarily metabolized or excreted unchanged in the bile. We suggest all MW>400 BDDCS Class 2, 3 and 4 compounds (but possibly excluding cations) since in the supplemental material for the Varma et al. [19] paper the authors report that 26% are neutral compounds, 60% are acids, 9% are zwitterions and 5% are cations. We note that we list olmesartan medoxomil as a BDDCS Class 1 drug [17]. This is not an error in classification since BDDCS designations are based on the drug formula (moiety) in the approved drug product. Olmesartan medoxomil is an ester prodrug that is metabolized to the acid form, which is then rate limited in its elimination by hepatic OATP uptake. In BDDCS, olmesartan as the acid form would be listed as a metabolite.

4.3 Renal versus Biliary Elimination of Unchanged Drug as the Major Excretion Pathway

The extended clearance methodologies listed above [19, 34–38] propose further usage of physicochemical and in silico characteristics such as molecular weight, lipophilicity, ionization state, and protein binding to further differentiate the major clearance mechanisms. We do not find these proposals to provide significant advantages over BDDCS. For example, we believe that ECCS [19] is too limited and has many more exceptions than BDDCS. For example, no drugs with MW>700 are considered, the system does not predict the importance of gut metabolism or disposition of prodrugs, ionization state is given more significance than justified and biliary excretion is not addressed except for drugs rate limited by hepatic uptake. We also find the list of compounds testing predictability in CPathPred [38] to be limited. As noted in Table 5, the in vitro model systems (determining permeability rate) perform quite well in differentiating extensive metabolism versus renal elimination and somewhat poorer, but acceptable in differentiating extensive metabolism versus biliary elimination, but are not successful at all in differentiating biliary versus renal elimination. Again, the in silico models (for permeability rate) are even less successful in differentiating metabolism from renal elimination and almost comparable to in vitro measures in differentiating metabolism and renal elimination from biliary elimination. Since transporters play a role in biliary elimination, we further suggest that these computed properties contribute little, if anything, in evaluating transporter effects.

We are concerned that the in silico methodologies being utilized have not adequately addressed biliary excretion of unchanged drug as an important disposition characteristic. Hosey et al. [39] has addressed this issue and proposed a classification scheme to differentiate renal and biliary elimination based on polarizability and predicted metabolic stability. We note that it is quite difficult to differentiate drug molecules that are primarily eliminated by metabolism from those that are excreted primarily via biliary excretion based only upon in silico characteristics. Here we present in Table 6, a listing of 20 drugs eliminated primarily by metabolism versus 20 drugs eliminated primarily via biliary excretion that cannot be differentiated based only on molecular weight, ionization state and CLogP. We believe that the proposed clearance concept methodologies [19, 34–38] need to include these 40 compounds in their validation tests. Reviewing the compounds utilized by Varma et al. [19] and by Toshimoto et al. [38] less than 10% of these drugs are included.

5.1 Metabolism as a Biowaiver Criterion

Since the BDDCS was proposed in 2005 [1], a number of extensions, providing new insights and predictions have been proposed. The excellent correlation between the high extent of absorption and the extent of metabolism led a number of experts in the field to recommend the use of BDDCS in classifying the permeability of marketed drugs using measures of the extent of metabolism following systemic absorption as justifying how much of the drug was absorbed [40]. This proposal was accepted by the EMA in their 2010 Guideline [41] and more recently by the FDA [8]. Ro5 does not incorporate criteria relevant to metabolism.

5.2 Food Effects

BDDCS also has a food effect extension. The exposure and “bioavailability of many drugs are greatly affected by concomitant food intake. Many factors are believed to contribute to these food effects, including changes in gastric emptying time, bile flow, pH of the intestine, splanchnic blood flow and gut wall metabolism. Different degrees of evidence support food effect on transporters”, as described by Custodio et al. [42], with Class 1 drugs generally showing no effect of high fat meals on the extent of absorption, Class 2 drugs generally showing an increase in the extent of absorption, and Class 3 drugs generally exhibiting a decrease in the extent of absorption. However, these general trends are only accurate approximately for 70% of drugs evaluated, and do not preclude the need to test the effect of food on the extent of absorption in humans for the to-be-marketed final dosage form [8, 42].

5.3 Uremic Toxins and Changes in Metabolism with Renal Failure

BDDCS extensions have also led to explanations related to drug disposition. Changes in the elimination of drugs where disposition is overwhelmingly due to metabolism were observed in a number of studies in patients with renal failure. Previously, this had been thought to be primarily due to product inhibition, i.e., the formed metabolite competes with the parent drug for metabolism, or the “effects of uremic toxins as either potential inhibitors or down regulators of metabolic enzymes. However, these hypotheses could be tested in vitro and were shown not to occur in many cases. We recognized that previously unexplained effects of renal disease on hepatic metabolism could result from accumulation of substances, such as uremic toxins, in renal failure that modify hepatic uptake and efflux transporters” [43–45]. “This mechanism could explain why BDDCS Class 2 drugs could demonstrate changes in metabolism in renal failure, whereas this would not be observed for BDDCS Class 1 drugs when in vitro uremic toxins did not alter microsomal metabolism. Inhibition of hepatic uptake by uremic toxins could also increase the exposure of BDDCS Class 3 and 4 drugs”, as we demonstrated with erythromycin [46].

5.4 Central Effects and Brain Efflux Transporters

BDDCS may also predict when central effects may or may not occur for new molecular entities. P-gp has the potential to modify brain concentrations and it has been hypothesized that for a drug to successfully penetrate and achieve pharmacodynamic effects in the brain, the drug should not be a P-gp substrate. However, we demonstrated that highly permeable-extensively metabolized, highly soluble (BDDCS Class 1) compounds can have a central effect at clinically approved doses, even if the drug is a good substrate for P-gp in vitro, regardless of whether the effect is desired or not [47]. We recently further addressed this issue examining the reliability of in vitro and in vivo methods for predicting the effect of P-gp on the delivery of antidepressants to the brain [28].

5.5 Flip-Flop Pharmaco*kinetics

As noted in Fig. 1, BDDCS Class 3 and 4 drugs require an uptake transporter in the intestine to achieve clinically meaningful systemic concentrations. For drugs with short disposition half-lives, the rate of this uptake process would be expected to be relatively slow. Thus, we suspected that drugs exhibiting flipflop kinetics (i.e., where the absorption half-life is longer than the elimination half-life) would predominantly be BDDCS Class 3 and 4 drugs, those requiring an uptake transporter, as we have recently shown [48].

5.6 Toxicity and Environmental Predictions

BDDCS extensions have begun to appear related to toxicity predictions and environmental implications. Broccatelli et al. [49] used the BDDCS to help consider for which drugs hERG voltage-gated potassium channel inhibition is likely to lead to Torsade de Pointes. Vuppalanchi et al. [50] and we [51] utilized BDDCS in evaluating drug induced liver injury, while our laboratory has also proposed the use of BDDCS to predict which anti-epileptic drugs will cause drug hypersensitivity reactions [52]. Finally, Daughton [53] has suggested that BDDCS could be used in attempts to decrease environmental exposure of active pharmaceutical ingredients.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1. Wu C-Y, Benet LZ. Predicting drug disposition via application of BCS: transport/absorption/elimination interplay and development of a biopharmaceutics drug disposition classification system. Pharm. Res. 2005;22:11–23. [PubMed] [Google Scholar]

2. Lipinski CA, Lombardo F, Dominy BW, Feeney PJ. Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Adv. Drug Deliv. Rev. 2001;46:3–26. [PubMed] [Google Scholar]

3. Lipinski CA. Drug-like properties and the causes of poor solubility and poor permeability. J. Pharmacology. Toxicol. Methods. 2000;44:235–249. [PubMed] [Google Scholar]

4. The International Transporter Consortium. Giacomini KM, Huang SM, Tweedie DJ, Benet LZ, Brouwer KL, Chu X, Dahlin A, Evers R, Fischer V, Hillgren KM, Hoffmaster KA, Ishikawa T, Keppler D, Kim RB, Lee CA, Niemi M, Polli JW, Sugiyama Y, Swaan PW, Ware JA, Wright SH, Wah Yee S, Zamek-Gliszczynski MJ, Zhang L. Membrane transporters in drug development. Nat. Rev. Drug Discov. 2010;9:215–236. [PMC free article] [PubMed] [Google Scholar]

5. Amidon GL, Lennernäs H, Shah VP, Crison JR. An analysis of the attrition of drug candidates from four major pharmaceutical companies. Nat. Rev. Drug Discov. 2015;14:475–486. [PubMed] [Google Scholar]

6. Oprea TI. Property distribution of drug-related chemical databases. J. Comput. Aided Mol. Des. 2000;14:251–264. [PubMed] [Google Scholar]

7. Waring MJ, Arrowsmith J, Leach AR, Leeson PD, Mandrell S, Owen RM, Pairaudeau G, Pennie WD, Pickett SD, Wang JB, Wallace O, Weir A. A theoretical basis for a biopharmaceutics drug classification: the correlation of in vitro drug product dissolution and in vivo bioavailability. Pharm. Res. 1995;12:413–420. [PubMed] [Google Scholar]

8. Food and Drug Administration. Guidance for industry: waiver of in vivo bioavailability and bioequivalence studies for immediate-release solid oral dosage forms based on a biopharmaceutics classification system. Rockville, MD: Food and Drug Administration; 2000. Available at http://www.fda.gov/cder/guidance/index.htm. [Google Scholar]

9. Wacher VJ, Wu CY, Benet LZ. Overlapping substrate specificities and tissue distribution of cytochrome P450 3A and P-glycoprotein: implications for drug delivery and activity in cancer chemotherapy. Mol. Carcinog. 1995;13:129–134. [PubMed] [Google Scholar]

10. Wu C-Y, Benet LZ, Hebert MF, Gupta SK, Rowland M, Gomez DY, Wacher VJ. Differentiation of absorption and first-pass gut and hepatic metabolism in humans: studies with cyclosporine. Clin. Pharmacol. Ther. 1995;58:492–497. [PubMed] [Google Scholar]

11. Cummins CL, Wu C-Y, Benet LZ. Sex-related differences in the clearance of Cytochrome P450 3A4 substrates may be caused by P-glycoprotein. Clin. Pharmacol. Ther. 2002;72:474–489. [PubMed] [Google Scholar]

12. Benet LZ, Cummins CL, Wu C-Y. Transporter-enzyme interactions: implications for predicting drug-drug interactions from in vitro data. Curr. Drug Metab. 2003;4:393–398. [PubMed] [Google Scholar]

13. Wu C-Y, Benet LZ. Disposition of tacrolimus in isolated perfused rat liver: influence of troleandomycin, cyclosporine, and GG918. Drug Metab. Dispos. 2003;31:1292–1295. [PubMed] [Google Scholar]

14. Lau YY, Wu C-Y, Okochi H, Benet LZ. Ex situ inhibition of hepatic uptake and efflux significantly changes metabolism: hepatic enzyme-transporter interplay. J. Pharmacol. Exp. Ther. 2004;308:1040–1045. [PubMed] [Google Scholar]

15. Benet LZ, Cummins CL, Wu C-Y. Unmasking the dynamic interplay between efflux transporters and metabolic enzymes. Int. J. Pharm. 2004;277:3–9. [PubMed] [Google Scholar]

16. Shugarts S, Benet LZ. The role of transporters in the pharmaco*kinetics of orally administered drugs. Pharm. Res. 2009;26:2039–2054. [PMC free article] [PubMed] [Google Scholar]

18. Hosey CM, Chan R, Benet LZ. BDDCS predictions, self-correcting aspects of BDDCS assignments, BDDCS assignment correction, and classification for more than 175 additional drugs. AAPS J. 2016;18:251–260. [PMC free article] [PubMed] [Google Scholar]

19. Varma MV, Styen SJ, Allerton C, El-Kattan AF. Predicting clearance mechanism in drug discovery: extended clearance classification system. Pharm. Res. 2015;32:3785–3802. [PubMed] [Google Scholar]

20. Niemi M, Pasan MK, Neuvonen PJ. SLCO1B1 polymorphism and sex affect the pharmaco*kinetics of pravastatin but not fluvastatin. Clin. Pharmacol. Ther. 2006;80:356–366. [PubMed] [Google Scholar]

21. Kalliokoski A, Niemi M. Impact of OATP transporters on pharmaco*kinetics. Br. J. Pharmacol. 2009;158:693–705. [PMC free article] [PubMed] [Google Scholar]

22. Cummins CL, Mangravite LM, Benet LZ. Characterizing the expression of CYP3A4 and efflux transporters (P-gp, MRP1, and MRP2) in CYP3A4-transfected Caco-2 cells after induction with sodium butyrate and the phorbol ester 12-O-tetradecanoylphorbol-13-acetate. Pharm. Res. 2001;18:1102–1109. [PubMed] [Google Scholar]

23. Cummins CL, Jacobsen W, Benet LZ. Unmasking the dynamic interplay between intestinal Pglycoprotein and CYP3A4. J. Pharmacol. Exp. Ther. 2002;300:1036–1045. [PubMed] [Google Scholar]

24. Cummins CL, Salphati L, Reid MJ, Benet LZ. In vivo modulation of intestinal CYP3A metabolism by P-glycoprotein: studies using the rat single-pass intestinal perfusion model. J. Pharmacol. Exp. Ther. 2003;305:306–314. [PubMed] [Google Scholar]

25. Cummins CL, Jacobsen W, Christians U, Benet LZ. CYP3A4-transfected Caco-2 cells as a tool for understanding biochemical absorption barriers: studies with sirolimus and midazolam. J. Pharmacol. Exp. Ther. 2004;308:143–155. [PubMed] [Google Scholar]

26. Benet LZ. The role of BCS (biopharmaceutics classification system) and BDDCS (biopharmaceutics drug disposition classification system) in drug development. J. Pharm. Sci. 2013;102:34–42. [PMC free article] [PubMed] [Google Scholar]

27. Broccatelli F, Cruciani G, Benet LZ, Oprea TI. BDDCS class prediction for new molecular entities. Mol. Pharmaceut. 2012;9:570–580. [PMC free article] [PubMed] [Google Scholar]

28. Zheng Y, Chen X, Benet LZ. Reliability of in vitro and in vivo methods for predicting the effect of P-glycoprotein on the delivery of antidepressants to the brain. Clin. Pharmaco*kinet. 2016;55:143–167. [PMC free article] [PubMed] [Google Scholar]

29. Al-Shawi MK, Omote H. The remarkable transport mechanism of P-glycoprotein; a multidrug transporter. J. Bioenerg. Biomembr. 2005;37:489–496. [PMC free article] [PubMed] [Google Scholar]

30. Leeson PD. Molecular inflation, attrition and the rule of five. Adv. Drug Deliv. Rev. 2016 Feb 1; [Epub ahead of print] [PubMed] [Google Scholar]

31. Hosey CM, Benet LZ. Predicting the extent of metabolism using in vitro permeability rate measurements and in silico permeability rate predictions. Mol. Pharmaceut. 2015;12:1456–1466. [PMC free article] [PubMed] [Google Scholar]

34. Camenisch G, Umehara K. Predicting human hepatic clearance from in vitro drug metabolism and transport data: a scientific and pharmaceutical perspective for assessing drug-drug interactions. Biopharm. Drug Dispos. 2012;33:179–194. [PubMed] [Google Scholar]

35. Kunze A, Poller B, Huwyler J, Camenisch G. Application of the extended clearance concept classification system (ECCCS) to predict the victim drug-drug interaction potential of statins. Drug Metabol. Personal Ther. 2015;30:175–188. [PubMed] [Google Scholar]

36. Camenisch G, Riede J, Kunze A, Huwyler J, Poller B, Umehara K. The extended clearance model and its use for the interpretation of hepatobiliary elimination data. ADMET DMPK. 2015;3:1–14. [Google Scholar]

37. Kusama M, Toshimoto K, Maeda K, Hirai Y, Imai S, Chiba K, Akiyama Y, Sugiyama Y. In silico classification of major clearance pathways of drugs with their physiochemical parameters. Drug Metab. Dispos. 2010;38:1362–1370. [PubMed] [Google Scholar]

38. Toshimoto K, Wakayama N, Kusama M, Maeda K, Sugiyama Y, Akiyama Y. In silico prediction of major drug clearance pathways by support vector machines with feature-selected descriptors. Drug Metab. Dispos. 2014;42:1811–1819. [PubMed] [Google Scholar]

39. Hosey CM, Broccatelli F, Benet LZ. Predicting when biliary excretion of parent drug is a major route of elimination in humans. AAPS J. 2014;16:1085–1096. [PMC free article] [PubMed] [Google Scholar]

40. Benet LZ, Amidon GL, Barends DM, Lennernäs H, Polli JE, Shah VP, Stavchansky SA, Yu LX. The use of BDDCS in classifying the permeability of marketed drugs. Pharm. Res. 2008;25:483–488. [PMC free article] [PubMed] [Google Scholar]

42. Custodio JM, Wu C-Y, Benet LZ. Predicting drug disposition, absorption/elimination/transporter interplay and the role of food on drug absorption. Adv. Drug Deliv. Rev. 2008;60:717–733. [PMC free article] [PubMed] [Google Scholar]

43. Sun H, Huang Y, Frassetto L, Benet LZ. Effects of uremic toxins on hepatic uptake and metabolism of erythromycin. Drug Metab. Dispos. 2004;32:1239–1246. [PubMed] [Google Scholar]

44. Sun H, Frassetto L, Benet LZ. Effects of renal failure on drug transport and metabolism. Pharmacol. Ther. 2006;109:1–11. [PubMed] [Google Scholar]

45. Reyes M, Benet LZ. Effects of uremic toxins on transport and metabolism of different biopharmaceutics drug disposition classification system xenobiotics. J. Pharm. Sci. 2011;100:3831–3842. [PMC free article] [PubMed] [Google Scholar]

46. Sun H, Frassetto LA, Benet LZ. Hepatic clearance, but not gut availability, of erythromycin is altered in patients with end-stage renal disease. Clin. Pharmacol. Ther. 2010;87:465–472. [PMC free article] [PubMed] [Google Scholar]

47. Broccatelli F, Larregieu CA, Cruciani G, Oprea TI, Benet LZ. Improving the prediction of the brain disposition for orally administered drugs using BDDCS. Adv. Drug Deliv. Rev. 2012;64:95–109. [PMC free article] [PubMed] [Google Scholar]

48. Garrison KL, Sahin S, Benet LZ. Few drugs display flip-flop pharmaco*kinetics and these are primarily associated with classes 3 and 4 of the BDDCS. J. Pharm. Sci. 2015;104:3229–3235. [PMC free article] [PubMed] [Google Scholar]

49. Broccatelli F, Mannhold R, Moriconi A, Giuli S, Carosati E. QSAR modeling and data mining link Torsades de Pointes risk to the interplay of extent of metabolism, active transport, and HERG liability. Mol. Pharmaceut. 2012;9:2290–2301. [PubMed] [Google Scholar]

50. Vuppalanchi R, Gotur R, Reddy KR, Fontana RJ, Ghabril M, Kosinski AS, Gu J, Serrano J, Chalasani N. Relationship Between characteristics of medications and drug-induced liver disease phenotype and outcome. Clin. Gastroenterol. Hepatol. 2014;12:1550–1555. [PMC free article] [PubMed] [Google Scholar]

52. Chan R, Wei C-y, Chen Y-t, Benet LZ. Use of the biopharmaceutics drug disposition classification system (BDDCS) to help predict the occurrence of idiosyncratic cutaneous adverse reaction associated with antiepileptic drug usage. AAPS J. 2016 Mar 7; [Epub Ahead of Print] [PMC free article] [PubMed] [Google Scholar]

53. Daughton CG. Eco-directed sustainable prescribing: feasibility for reducing water contamination by drugs. Sci. Total Environ. 2014;15:392–404. [PubMed] [Google Scholar]

54. Vaidyanathan S, Jarugula V, Dieterich HA, Howard D, Dole WP. Clinical pharmaco*kinetics and pharmacodynamics of aliskiren. Clin. Pharmaco*kinet. 2008;47:515–531. [PubMed] [Google Scholar]

55. Rolan PE, Mercer AJ, Tate E, Benjamin I, Posner J. Disposition of atovaquone in humans. Antimicrob. Agents Chemother. 1997;41:1319–1321. [PMC free article] [PubMed] [Google Scholar]

56. Singlas E. Clinical pharmaco*kinetics of azithromycin. Pathol Biol. 1995;43:505–511. [PubMed] [Google Scholar]

57. Arvidsson A, Alvan G, Angelin B, Borga O, Nord CE. Ceftriaxone: renal and biliary excretion and effect on the colon microflora. J. Antimicrob. Chemother. 1982;10:207–215. [PubMed] [Google Scholar]

58. Hitzenberger G, Takacs F, Pittner H. Pharmaco*kinetics of the beta-adrenergic receptor blocking agent celiprolol after single intravenous and oral administrations in man. Arzneimittelforschung. 1983;33:50–52. [Google Scholar]

59. Rudi J, Raedsch R, Gerteis C, Schlenker T, Plachky J, Walter-Sack I, Sabourad A, Scherrmann JM, Kommerell B. Plasma kinetics and biliary excretion of colchicine in patients with chronic liver disease after oral administration of a single dose and after long-term treatment. Scand. J. Gastroenterol. 1994;29:346–361. [PubMed] [Google Scholar]

60. Storstein L. Studies on digitalis. III. Biliary excretion and enterohepatic circulation of digitoxin and its cardioactive metabolites Clin. Pharmacol. Ther. 1975;17:313–320. [PubMed] [Google Scholar]

61. Gleiter CH, Morike KE. Clinical pharmaco*kinetics of candesartan. Clin Pharmaco*kinet. 2002;41:7–17. [PubMed] [Google Scholar]

62. Yang X, Gandhi YA, B D, Morris ME. Prediction of biliary excretion in rats and humans using molecular weight and quantitative structure pharmaco*kinetic relationships. AAPS J. 2009;11:511–525. [PMC free article] [PubMed] [Google Scholar]

63. Ayrton A, Morgan P. Role of transport proteins in drug absorption, distribution and excretion. Xenobiotica. 2001;31:469–497. [PubMed] [Google Scholar]

64. Swift B, Tian X, Brouwer KLR. Integration of preclinical and clinical data with pharmaco*kinetic modeling and simulation to evaluate fexofenadine as a probe for hepatobiliary transport function. Pharm. Res. 2009;26:1942–1951. [PMC free article] [PubMed] [Google Scholar]

65. Lippert C, Ling J, Brown P, Burmaster S, Eller M, Cheng L, Thompson R, Weir S. Mass balance and pharmaco*kinetics of MDL 16,455A in healthy, male volunteers. Pharm. Res. 1995;12:S-390. [Google Scholar]

66. Schwab D, Grauer M, Hahn EG, Muhldorfer S. Biliary secretion of moxifloxacin in obstructive cholangitis and the non-obstructed biliary tract. Ailment Pharmacol. Ther. 2005;22:417–422. [PubMed] [Google Scholar]

68. Lee SP, Paxton JW, Choong YS. Plama and biliary disposition of pirenzepine in man. Clin. Exp. Pharm. Phys. 1986;13:241–248. [PubMed] [Google Scholar]

69. Verardi S, Verardi V. Bile rifaximin concentration after oral administration in patients undergoing cholecystectomy. Farmaco. 1990;45:131–135. [PubMed] [Google Scholar]

70. Hong KS. Rifaximin for the treatment of acute infectious diarrhea. Therap. Adv. Gastroenterol. 2011;4:227–235. [PMC free article] [PubMed] [Google Scholar]

71. Puri SK, Lassman HB. Roxithromycin: a pharmaco*kinetic review of a macrolide. J. Antimicrob. Chemother. 1987;20(Suppl B):89–100. [PubMed] [Google Scholar]