macropowerful

1. Bcs Classification Of Drugs
2. Bcs Classification System Database

Classification of APIs according to the BCS According to the HHS-FDA definitions in the documents cited above, the four different categories possible for an API according to the BCS are.

• Mar 9, 2006 - Dear friends Can someone provide information about existing BCS class databases or drug permeability databases?
• Since the biopharmaceutics classification system (BCS) was introduced in 1995, it has had an increasing impact on regulatory practice. The BCS presented a new paradigm in bioequivalence, based on scientific principles.

Associated Data

Abstract

Here, we compile the Biopharmaceutics Drug Disposition Classification System (BDDCS) classification for 927 drugs, which include 30 active metabolites. Of the 897 parent drugs, 78.8% (707) are administered orally. Where the lowest measured solubility is found, this value is reported for 72.7% (513) of these orally administered drugs and a dose number is recorded. The measured values are reported for percent excreted unchanged in urine, LogP, and LogD_7.4 when available. For all 927 compounds, the in silico parameters for predicted Log solubility in water, calculated LogP, polar surface area, and the number of hydrogen bond acceptors and hydrogen bond donors for the active moiety are also provided, thereby allowing comparison analyses for both in silico and experimentally measured values. We discuss the potential use of BDDCS to estimate the disposition characteristics of novel chemicals (new molecular entities) in the early stages of drug discovery and development. Transporter effects in the intestine and the liver are not clinically relevant for BDDCS class 1 drugs, but potentially can have a high impact for class 2 (efflux in the gut, and efflux and uptake in the liver) and class 3 (uptake and efflux in both gut and liver) drugs. A combination of high dose and low solubility is likely to cause BDDCS class 4 to be underpopulated in terms of approved drugs (N = 53 compared with over 200 each in classes 1–3). The influence of several measured and in silico parameters in the process of BDDCS category assignment is discussed in detail.

Electronic supplementary material

The online version of this article (doi:10.1208/s12248-011-9290-9) contains supplementary material, which is available to authorized users.

In 2005, Wu and Benet () introduced the Biopharmaceutics Drug Disposition Classification System (BDDCS). Wu and Benet recognized that there was a very strong correlation between the intestinal permeability rate and the extent of metabolism. For example, Benet et al. () noted that for the 29 drugs and endogenous substances for which human jejunal permeability rate measurements were available, there was an excellent correlation between these permeability rate measurements and the extent of drug metabolism in humans. Fourteen of the 16 drugs exhibiting human intestinal permeability rates greater than metoprolol were extensively metabolized, while 11 of 12 drugs showing permeability rates less than metoprolol were poorly metabolized. Two drugs showing disparity between the permeability rate and metabolism, cephalexin and losartan, exhibit permeability rates that differ by no more than 16% from metoprolol (). Since the coefficients of variation for the human permeability parameters range from 29% to 130%, these borderline compounds may in fact also have followed the correlation. The correlation between the extent of metabolism and human intestinal jejunal permeability was markedly better than that observed for intestinal jejunal permeability and partition coefficient by Takagi et al. (), who noted that Log P measured and calculated correctly predict high versus low permeability only about two thirds of the time. Wu and Benet () reasoned that it might be easier to utilize metabolism in assigning drug classification since it is difficult and expensive to determine human intestinal permeabilities and since it is also difficult to obtain quantitative mass balance measures that show ≥90% absorption, the FDA criterion for a biowaiver as defined in the FDA BCS Guidance (4), based on the work of Amidon et al. (). Therefore, in proposing the BDDCS classification system, Wu and Benet () substituted extensive and poor metabolism for high and low permeability in the BCS while utilizing the same criteria as the FDA for high and low solubility. That is, a high solubility compound at the highest marketed dose strength would be soluble in 250 mL of water over the pH range of 1–7.5 at 37°C. Using the BDDCS, Wu and Benet () classified 168 drugs based on the extent of metabolism and solubility.

BDDCS VERSUS BCS

Although BDDCS grew out of the FDA’s BCS Guidance (4), Wu and Benet () proposed BDDCS as a means to predict the drug disposition characteristics of novel chemicals (here, referred to as “new molecular entities”, NMEs) during the early stages of drug discovery and development. Such examples will be discussed below. Recently, Benet and Larregieu () reviewed the differences between BCS and BDDCS in terms of purpose and basis. The purpose of BCS is to facilitate biowaivers of in vivo bioequivalence studies for drugs that exhibit no significant intestinal absorption problems. In contrast, the purpose of BDDCS is to predict the drug disposition of NMEs as well as potential drug–drug interactions for NMEs and drugs on the market with respect to the intestine and liver. Very recently, a consensus paper with respect to BCS, BDDCS, and regulatory guidances has been published ().

Both BCS and BDDCS use the same criteria for solubility. Therefore, there is no difference in the basis between the two systems with respect to this parameter. As noted above, BDDCS predictions and classification are based on the intestinal permeability rate, not the extent of permeability. There is some ambiguity with respect to the basis for BCS, as reviewed by Benet and Larregieu (). The initial permeability studies of Amidon, Lennernäs, and colleagues (,), as summarized by Takagi et al. (), show a good correlation between human intestinal permeability rate and the extent of absorption, as detailed earlier in the first paragraph of this paper. However, the criterion listed in the FDA BCS Guidance (4) is “…a drug substance is considered to be highly permeable when the extent of absorption in humans is determined to be 90% or more of an administered dose based on a mass balance determination or in comparison to an intravenous reference dose” [emphasis added by the FDA]. Although permeability rate methods are listed in the FDA BCS Guidance (4), we are unaware of any drug that has been certified by the FDA as class 1 eligible for in vivo biowaiver where there is no confirmatory ≥90% absorption data. This ambiguity does not exist in the Guideline on the Investigation of Bioequivalence issued in 2010 by the European Medicines Agency (EMA), which only allows in vivo biowaivers based on the extent of absorption (9). This difference in the permeability basis between BCS and BDDCS is brought home in a recent publication by FDA scientists (). Chen and Yu () note that the FDA has classified as “highly permeable” a number of drugs where absorption is ≥90% in humans, but the measured permeability rates of these compounds are less than that for metoprolol (cefadroxil, cephradine, levofloxacin, loracarbef, ofloxacin, and sotalol), and in one case (pregabalin), the measured permeability rate is less than that for mannitol. As previously recognized (), BCS is influenced by transporter effects. For example, large amino acid transporter-1 (LAT-1) is expressed in Caco-2 cells (), and pregabalin is a LAT-1 substrate as noted in the package insert (12), which may explain this discrepancy. Since pregabalin is a zwitterion, its high oral bioavailability (≥90%) may be attributed to LAT-1 transport, an effect that is not taken into account by BCS. Thus, although in general drugs exhibiting high intestinal permeability rates show a high extent of absorption and a high extent of metabolism in both BCS and BDDCS, there are a number of drugs that could be classified as highly permeable in the BCS system based on absorption ≥90% but would be predicted to be poorly metabolized based on the low intestinal permeability rate, the basis for BDDCS classification. By evaluating metabolism, not permeability, BDDCS is not subject to variability due to transporter effects.

In general, classification of drugs between BCS and BDDCS only differ about 5–10%. However, for class 1 drugs where FDA has granted biowaivers, we estimate that the difference between BCS and BDDCS occurs for about 40% of drugs. We cannot make a more accurate estimate since the listing of all drugs granted biowaivers by the FDA is confidential. The percentage difference is high due to the ease in determining whether a drug is >90% absorbed (class 1 in BCS) when a drug is almost completely eliminated unchanged in the urine (class 3 in BDDCS).

BDDCS AND ITS USE

In 1995, Wu and Benet () reviewed 131 drugs that had been classified into the four BCS categories in the literature through the end of 1994. Ten of these drugs had been listed in different classes by different authors. Wu and Benet () recognized that the major route of elimination in humans for the great majority of high-permeability class 1 and class 2 drugs was metabolism, while the major route of elimination for the poorly permeable class 3 and class 4 drugs in humans was renal and biliary excretion of unchanged drug. They also noted that the major route of elimination via cytochrome P450 3A4 (CYP3A4) was only observed for the class 1 and class 2 drugs and that for the class 3 and class 4 drugs CYP3A4 was not a major contributor to elimination for any. Since the extent of metabolism is better characterized than the extent of absorption, for marketed drugs, Wu and Benet proposed that in BDDCS, drugs be categorized in terms of the extent of metabolism and solubility versus permeability rate and solubility (). This immediately eliminated the situation where drugs were classified in more than one class because of the uncertainty of permeability measures from study to study. The implication from the BDDCS for an NME is that if a surrogate measure of intestinal absorption rate is available, such as permeability rate through a Caco-2 cellular system, it would be possible to predict the major route of elimination for this new molecular entity in humans prior to its in vivo dosing to either animals or humans. Work is ongoing in our laboratory to determine the degree of accuracy in predicting BDDCS class for an NME based only on in vitro permeability measures prior to studies determining the extent of metabolism. Thus, Benet and Wu () proposed the BDDCS as shown in Fig. 1 with ≥70% metabolism being the cutoff for extensive metabolism. They also noted that there were relatively few drugs where the extent of metabolism was between 30% and 70% and that most drugs are either very highly metabolized or very poorly metabolized.

The Biopharmaceutics Drug Disposition Classification System (BDDCS) as proposed by Wu and Benet ()

Figure 2 summarizes the predictions from BDDCS related to the effects of enzymes and transporters in the gut and liver following oral dosing of drugs (,). For class 1, highly soluble—high permeability rate—extensively metabolized drugs, transporter effects in the intestine and the liver have no clinical impact. Even compounds like verapamil, which can be shown in certain cellular systems (e.g., MDR1-MDCK) to be a substrate for P-glycoprotein (P-gp), exhibit no clinically significant P-gp substrate effects in the gut and the liver. Thus, a major proposition of BDDCS is that although class 1 drugs may be shown in cellular systems to be substrates for transporters found in the intestine and the liver, this has no clinical relevance. However, a caution is in order here. At this time, BDDCS predictions only apply to the intestine and the liver since class 1 drugs could be substrates for transporters at the blood–brain barrier and in the kidney.

Transporter effects predicted by BDDCS following oral dosing

From Fig. 2, it can be seen that for class 2 drugs, efflux transporter effects will predominate in the intestines. Thus, transporter–enzyme interplay will be primarily important for class 2 compounds that are substrates for CYP3A and phase II gut enzymes (e.g., glucuronosyltransferases, sulfotransferases), where efflux transporter effects can control the access of the drug to the gut enzymes. BDDCS predicts that both uptake and efflux transporters can affect class 2 drug disposition in the liver. Thus, inhibition or induction of uptake hepatic transporters such as the SLCOs (OATPs) and the SLC22As (OATs and OCTs) as well as the drug efflux hepatic transporters ABCB1 (P-gp), ABCG2 (BCRP), and ABCCs (MRPs) can lead to changes in hepatic metabolism even when hepatic enzymes are unaffected.

BDDCS predicts that for class 3, highly soluble—poor permeability rate—poorly metabolized drugs, uptake transporters will be important for intestinal absorption and liver entry for these poorly permeable compounds (Fig. 2). However, once these drugs get into the enterocyte or the hepatocyte, efflux transporter effects can also occur. Similarly, uptake and efflux transporter effects would be expected for the poorly soluble class 4 compounds. Because they are numerically underrepresented, one might expect that class 4 drugs are more difficult to manage therapeutically, i.e., there are transporter effects just as in class 2, except these drugs are not significantly metabolized. As will be shown subsequently, plotting the maximum recommended therapeutic daily dose () versus BDDCS and looking at the “actives” (low dose) versus “inactives” (high dose, safer) revealed no such relationship. It is more likely that fewer drugs are represented in class 4 because of the combined negative characteristics of high dose and (comparatively) low solubility, which leads to high variability. Since the FDA criteria for solubility is measured in water, we suspect that the approved class 4 drugs have adequate solubility in the natural surfactant containing intestinal fluids.

Details and explications for the predictions in Fig. 2 have been presented in recent reviews from the Benet lab (,,). However, in large part, the predictions in Fig. 2 were based on clinical and experimental observations. That is, we are unaware of any clinically significant effects of the uptake and efflux transporters in the gut and liver on class 1 drugs, even when such drugs have been shown in cellular systems or other organs besides the gut and the liver to be substrates of the uptake and efflux transporters. These effects, however, might become relevant in overdosage situations, e.g., when combined with strong inhibitors of their respective metabolizing enzyme, when liver failure is manifest, or when accidentally overdosed. Similarly, we are unaware of the clinically significant effects of the uptake transporters in the gut for class 2 drugs even when these drugs are shown to be substrates of uptake transporters in the liver.

BDDCS also allows potential drug–drug interactions to be predicted (,). For class 1 drugs, only metabolic interactions need to be considered in the intestine and the liver. For class 2 drugs, metabolic, efflux transporter, and efflux transporter–enzyme interplay in the intestine must be taken into consideration, while in the liver, metabolic, uptake transporter, efflux transporter, and transporter–enzyme interplay (both uptake and efflux) can occur. For class 3 and class 4 drugs, uptake transporter, efflux transporter, and uptake–efflux transporter interplay will be of major importance.

BDDCS classification may also be useful in predicting the effect of high-fat meals on the extent of bioavailability (F) for an NME. In general, F for class 1 drugs is unaffected by high-fat meals; F is generally increased for class 2 drugs and generally decreased for class 3 drugs (). Custodio et al. () have observed that these findings would be the outcomes expected if a component of high-fat meals inhibited both the uptake and efflux transporters. However, even if this were true, high-fat meals would be expected to have many other effects than inhibiting transporters. We estimate that the predicted effect of high-fat meals on F is only correct about 70% of the time.

CAUTION

It is surprising that such a simple four-category process as indicated in Fig. 2 works so well in predicting drug disposition, transporter–enzyme effects, and drug interactions. It is obvious, however, that this simple four-category system will not predict every interaction. BDDCS does not propose that every drug in the class will be substrates or not substrates for the uptake and efflux transporters. Rather, BDDCS helps prioritize what interactions should and should not be investigated. For example, the class 2 drug felodipine has been shown not to be affected by the intestinal or hepatic efflux transporters (). Recently, our laboratory has shown the importance in humans of hepatic uptake transporters for the drugs atorvastatin and glyburide (,). These interactions were predicted based on cellular, isolated organ, and animal studies (–). Even when such preliminary studies confirm BDDCS predictions, this may not always be the case. Warfarin is a class 2 drug and, thus according to BDDCS, may be a substrate for a hepatic uptake transporter. In vitro studies in human and rat hepatocytes showed that rifampin would decrease warfarin metabolism by 30% (), a similar extent to that found for our in vitro results with glyburide (). However, our recently published study examining the effects of a single dose of rifampin on the pharmacokinetics of warfarin in healthy volunteers showed that OATP uptake in vivo in humans was not clinically significant for warfarin (). Similarly, although warfarin appears to be both a substrate and inhibitor of liver-bound P-gp (26), this has not been regarded as clinically significant; one P-gp haplotype, however, is clearly associated with low-dose warfarin in a 201 patient sample (). This emphasizes again the caution that BDDCS only predicts with respect to transporters what might occur, but not that the effect will always occur. Furthermore, our example above () reinforces the well-recognized concept that observations in cellular systems and animal models must be tested in vivo in humans before the significance of the effect is assumed.

WHY DID WE PREPARE THIS PAPER?

Wu and Benet () recognized that the FDA’s BCS approach (4) held the potential for predicting the drug disposition characteristics and drug interactions for NMEs as well as for drugs on the market. The use of BDDCS in the area of systems chemical biology () has been previously outlined (), and computational models to assign BDDCS class from molecular structure have been proposed (). However, to test the usefulness of BDDCS, to examine patterns within the BDDCS classes and among them, and to gain further perspectives, it is necessary to compile a large database, at least with respect to drugs that have reached the market. Since BDDCS makes predictions related to hepatic elimination in addition to intestinal absorption, such a database should include as many drugs as possible where systemic concentrations are relevant. Thus, approximately one quarter of the drugs categorized here are administered exclusively by non-oral routes. We also felt strongly that the information provided in the database should be based, where available, not only on the in silico predictions of those parameters but also on experimental values. We noted that many of the solubility values used in BCS analyses are frequently in silico predictions of solubility. For example, in the often quoted paper of Willmann et al. () describing a physiological model for the estimation of the fraction dose absorbed in humans, the measured solubility values were only included for only 22 of the 126 drugs evaluated. The solubility for the great majority of the drugs utilized in the Willmann et al. () analysis came from the compilation of Zhao et al. (). These latter workers evaluated human intestinal absorption data for 241 drugs (). Of the 241 drugs, Zhao et al. compared the experimental results for 26 of the compounds with the predicted solubility utilizing the method of Meylen et al. (33) based on octanol water partition coefficients. For these 26 drugs, the measured versus calculated solubility differs by a factor of 5.7 ± 6.0-fold, the greatest difference being 23-fold. Thus, at present, in silico methodologies for aqueous solubility prediction are not sufficiently accurate for the BDDCS analysis not only due to the inherent limitations of such methods but also to its definition, i.e., BDDCS solubility categorization depends on the maximum strength dose and the effect of pH. Experimental solubility values were included in this compilation wherever they could be found in the literature (577 drugs). Qualitative evaluations such as “practically insoluble in water,” which relate to upper limits, and “highly soluble in water” were used in the absence of published solubility values from a reliable source and were the basis of the BDDCS assignment when no measured value is listed in the table.

Frequently, large compilations such as the ones presented here are carried out with the assistance of graduate students, postdoctoral fellows, and auxiliary personnel. In our experience, this can lead to unevenness in the quality of the data presented in the table. For each of the drugs listed here, decisions concerning the experimental values to be listed and classification assigned were made during the multiple joint meetings of Drs. Benet and Oprea between February 2007 and February 2011. Then, Dr. Broccatelli captured potential errors by cross-checking many references. Therefore, if there are errors in the parameters or in the assignments, this can only be attributable to the authors.

MEASURED PARAMETERS (IN ORDER OF DIFFICULTY)

Solubility

There are a number of issues concerning the choice of the high versus low solubility criteria and which representative experimental values should be listed. The high solubility criterion that the highest dose strength on the market is soluble in 250 mL or less of water over the pH range 1–7.5 at 37°C was an arbitrary decision made by Amidon et al. () and incorporated into the regulatory Guidances (4,9). Wu and Benet () found that this cutoff criterion in BCS appeared to work well for BDDCS, and thus, we have continued to use this arbitrary, discriminatory criterion. The FDA criteria (4) require the solubility measurements to be made in water, not simulated intestinal fluid containing a surfactant, and the solubility values listed in the table are values in water. Furthermore, the FDA criteria evaluate the cutoff between high and low solubility using the value for the lowest solubility over the pH range 1–7.5 (realistically measured at pH 1.2, 4.5, and 6.8 as indicated in the FDA Guidance). Furthermore, the solubility is to be measured at 37°C. The values in Tables I, ​,II,II, III, ​,IV,IV, and ​andVV are the authors’ best recommendation based on experimental literature data for the lowest solubility under the conditions listed above. The solubility criteria over the pH range 1–7.5 can create marked differences from compilation to compilation for drugs that are salts. For example, one may find in the package insert for atazanavir sulfate that the drug “is slightly soluble in water (4–5 mg/mL, free base equivalent) with the pH of a saturated solution in water being about 1.9.” Reporting this solubility would lead to the assignment of atazanavir sulfate as a class 1 drug in BCS and BDDCS. However, it is known that atazanavir, as for almost all the protease inhibitors, exhibits very poor solubility in solutions at pH above 5.5, although these solubilities are not reported. Therefore, we list atazanavir sulfate as a class 2 drug and no specific solubility is listed in Table II. Another example of a basic drug is imatinib mesylate (and other kinase inhibitors) where the package insert states “Imatinib mesylate is soluble in aqueous buffers ≤pH 5.5 [a published value of 200 mg/mL can be found] but is very slightly soluble to insoluble in neutral/alkaline aqueous buffers.” Therefore, we list this drug in Table II as class 2 with a solubility of 1 mg/mL, determined at pH 7.4. Similar variances can occur for acidic drugs where solubility may be very low at pH 1. This concern that the pH range maybe too restrictive has been addressed () and explicated () previously for acidic drugs. Individual references are not given for the solubility values because in many cases the values in the tables are the authors’ consensus view of the appropriate value based on a number of experimental results as discussed above. This is a technique followed by Dr. Benet when he first introduced the table of pharmacokinetic parameters in Goodman and Gilman and represents a decision of the two senior authors alone in reviewing all of the literature data that they could find. A single solubility value is given in the five tables in milligrams per milliliter.

Table I

Measured and In Silico Data for 351 BDDCS Class 1 Drugs

Table II

Measured and In Silico Data for 265 BDDCS Class 2 Drugs