European Journal of Pharmaceutical Sciences 57 (2014) 224–231

Contents lists available at ScienceDirect

European Journal of Pharmaceutical Sciences

journal homepage: www.elsevier.com/locate/ejps

Is the full potential of the biopharmaceutics classification system

reached?
Christel A.S. Bergström a,⇑, Sara B.E. Andersson a, Jonas H. Fagerberg a, Gert Ragnarsson b, Anders Lindahl b
a
Department of Pharmacy, Uppsala University, Biomedical Centre, P.O. Box 580, SE-751 23 Uppsala, Sweden
b
Medical Products Agency, Scientific Expertise, Box 26, SE-751 03 Uppsala, Sweden

a r t i c l e i n f o a b s t r a c t

Article history: In this paper we analyse how the biopharmaceutics classification system (BCS) has been used to date. A
Received 3 June 2013 survey of the literature resulted in a compilation of 242 compounds for which BCS classes were reported.
Received in revised form 5 September 2013 Of these, 183 compounds had been reported to belong to one specific BCS class whereas 59 compounds
Accepted 15 September 2013
had been assigned to multiple BCS classes in different papers. Interestingly, a majority of the BCS class 2
Available online 25 September 2013
compounds had fraction absorbed (FA) values >85%, indicating that they were completely absorbed after
oral administration. Solubility was computationally predicted at pH 6.8 for BCS class 2 compounds to
Keywords:
explore the impact of the pH of the small intestine, where most of the absorption occurs, on the solubility.
Biopharmaceutics classification system
Solubility
In addition, the solubilization capacity of lipid aggregates naturally present in the intestine was studied
Biorelevant dissolution computationally and experimentally for a subset of 12 compounds. It was found that all acidic com-
Dose number pounds with FA > 85% were completely dissolved in the pH of the small intestine. Further, lipids at the
Poorly soluble concentration used in fasted state simulated intestinal fluid (FaSSIF) dissolved the complete dose given
In silico of the most lipophilic (log D6.5 > 3) compounds studied. Overall, biorelevant dissolution media (pure buf-
fer of intestinal pH or FaSSIF) identified that for 20 of the 29 BCS class 2 compounds with FA > 85% the
complete dose given orally would be dissolved. These results indicate that a more relevant pH restriction
for acids and/or dissolution medium with lipids present better forecast solubility-limited absorption
in vivo than the presently used BCS solubility criterion. The analysis presented herein further strengthens
the discussion on the requirement of more physiologically relevant dissolution media for the in vitro sol-
ubility classification performed to reach the full potential of the BCS.
Ó 2013 Elsevier B.V. All rights reserved.

1. Introduction BDDCS uses metabolism instead of permeability in its classifica-

tion. It is proposed to be used to identify disposition and drug–drug
The biopharmaceutics classification system (BCS) categorizes interaction patterns primarily in the intestine and liver (Wu and
drug molecules on the basis of their permeability and solubility Benet, 2005; Benet, 2013). In contrast, the BCS and DCS have the
(Amidon et al., 1995), and serves as a tool to identify compounds same permeability definition, but the DCS is based on larger vol-
eligible for a biowaiver of in vivo bioequivalence (FDA, 2000; umes for the solubility definition. Butler and Dressman proposed
WHO, 2006; EMA, 2010) The BCS has gained much attention be- that 500 mL, rather than the 250 mL used in the BCS definition,
cause of its potential to reduce the need for clinical studies during better reflects volumes in the small intestine (Butler and
e.g. life cycle management and production of generic compounds. Dressman, 2010). At the same time, the larger volume (500 mL)
It is also used to identify potential absorption problems after oral is believed to compensate for the disappearance of drug from the
administration. The BCS has evolved in many different directions. intestine (i.e. the permeability) during drug dissolution and solu-
As an example, it is used to signal need for formulation design bility assessment. The DCS is particularly useful for indicating
(BCS class 2/4) and/or chemical modifications (BCS class 3/4) to im- whether BCS class 2 compounds are dissolution-rate limited or
prove absorption (Pouton, 2006). It has also resulted in the devel- solubility limited; these are important factors for guiding further
opment of the Biopharmaceutics Drug Disposition Classification formulation design.
System (BDDCS) (Wu and Benet, 2005) and the Developability From a regulatory perspective the BCS provides a scientific
Classification System (DCS) (Butler and Dressman, 2010). The BCS framework to determine whether bioequivalence (BE) studies are
and BDDCS are based on the same solubility definition, but the needed for new formulations of drugs or if in vitro data suggest a
biowaiver to be designated. BE is defined as the compound having
⇑ Corresponding author. Tel.: +46 18 471 4118; fax: +46 18 471 4223. an AUC and Cmax in the interval of 80–125% of the reference prod-
E-mail address: [email protected] (C.A.S. Bergström).
uct. The 90% confidence interval for the test/reference ratio should

0928-0987/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.ejps.2013.09.010
 C.A.S. Bergström et al. / European Journal of Pharmaceutical Sciences 57 (2014) 224–231 225

be contained within this range (EMA, 2010). Attempts have been (Dressman et al., 2001). It is well-known that lipophilic drug mol-
made to calculate the economic benefits of BCS biowaivers (Cook ecules may gain extensively in solubility when dissolved in more
and Bockbrader, 2002; Cook et al., 2010). It is difficult to perform physiologically relevant dissolution media such as fasted or fed
such economical calculations because of assumptions such as the state simulated intestinal fluids (FaSSIF and FeSSIF, respectively)
frequency of compounds in each BCS class and the number of sub- (Mithani et al., 1996; Fagerberg et al., 2010; Soderlind et al.,
jects needed to prove BE (which is related to the variability in AUC 2010; Fagerberg et al., 2012), and that the obtained solubility in
and Cmax expected for a drug). The economic impact is also higher these simulated fluids better reflect the solubility measured in hu-
in Europe as compared to the U.S. due to the EMA extension of man intestinal fluids for such compounds (Nicolaides et al., 2001;
biowaivers to BCS class 3 drug products. However, taking into ac- Soderlind et al., 2010; Clarysse et al., 2011). Hence, the use of sim-
count both BCS class 1 and class 3 biowaivers, it is estimated that ple buffers instead of biorelevant dissolution media may result in a
128–147 million dollars could be saved each year in reduced clin- significant underestimation of the concentration reached in the
ical study costs, out of which approximately 66–76 million dollars intestinal fluid. The solubility definition of the BCS has been vigor-
reflect saved costs for BCS class 1 (Cook et al., 2010). Other benefits ously discussed in particular with regard to pH and lipid content
of increased use of the BCS-based biowaivers are that unnecessary but also from the volume perspective (Dressman et al., 2001; Yu
testing in humans is avoided and that the new products reach the et al., 2002; Rinaki et al., 2003; Polli et al., 2004; Rinaki et al.,
market faster. Despite these obvious advantages, pharmaceutical 2004; Yazdanian et al., 2004; Lennernas and Abrahamsson, 2005;
companies appear reluctant to pursue BCS-based biowaivers. In Polli et al., 2008; Davit et al., 2012; Varma et al., 2012). In this
2010 (i.e. 10 years after the FDA introduced the guidelines for study we explored the current solubility definition of BCS by study-
BCS-based biowaivers), Cook et al. reported that the number of ing 242 compounds for which the BCS class could be extracted
applications for biowaivers had started to increase; nevertheless, from the literature. Of these, focus was set on the 58 BCS class 2
the FDA had only designated 27 compounds a BCS class 1 biowai- compounds for which the impact of pH and lipids present in the
ver by this year. In a recent analysis at the Swedish Medical Prod- fasted state intestinal fluid on the observed solubility was investi-
ucts Agency, it was revealed that only 10% (21 of 208) and 17% (28 gated and analysed with an in vivo FA perspective.
of 167) of the oral solid drug products (immediate release) had ap-
plied for a BCS-based biowaiver in 2011 and 2012, respectively.
2. Materials and methods
These are low percentages, given that approximately two-thirds
of marketed compounds are BCS class 1 or 3, (Takagi et al., 2006;
2.1. Dataset and characteristics
Cook et al., 2010) for which the companies can file for biowaivers
in Europe. From the above presented analysis it becomes clear that
A set of compounds (n = 242) was extracted from published
the pharmaceutical industry currently does not take full advantage
datasets for which BCS classifications existed (Tables 1 and 2)
of the BCS. One reason may be that not all regulatory authorities
(Bergstrom et al., 2003; Lindenberg et al., 2004; Yazdanian et al.,
approve biowaivers based on the BCS. Although the pharmaceuti-
2004; Wu and Benet, 2005; Yang et al., 2007; Butler and Dressman,
cal industry is global, intending to target all markets so as to fully
2010; Kleberg et al., 2010; Ramirez et al., 2010). Thereafter, we
exploit the investments they have made in their drug products, the
focused on BCS class 2 compounds (n = 58). For these substances,
regulatory authorities are not equally globalized. For instance, the
molecular descriptors and e.g. pKa, log P and log D6.5 were calcu-
EMA and the FDA approve biowaivers based on the BCS (although
lated with the software ADMET Predictor (SimulationsPlus, CA).
using slightly different criteria), but Japan does not.
Fraction absorbed (FA) values were taken from the literature
The current BCS hence seems not to be fully exploited in that
(Table 3). If FA data were not available, oral bioavailability (F)
the number of biowaivers designated still is low. However, there
was used as a surrogate for those BCS class 2 compounds with
may also be room for increased usage of BCS-based biowaivers
F > 85%. Our decision was based on the EMA guideline which uses
by changing the in vitro profiles required for active pharmaceutical
a FA of 85% as the high permeability cut-off (EMA, 2010). Further-
ingredients (APIs) to be defined as BCS class 1 compounds. The cur-
more, if F is 0.85 or more, then the FA must be 85% or greater.
rent definition for solubility has, in addition to the amendments
made through the DCS system, been subject to debate. The pH-
interval introduced by the FDA has been discussed, because of 2.2. Materials
the relatively high pH of 7.5 used as the upper limit of the pH inter-
val in which the dose needs to be soluble, and as a response to this Carbamazepine, dapsone, lansoprazole, nevirapine, ofloxacin,
discussion the EMA guideline uses a pH interval from 1 to 6.8. The phenazopyridine, praziquantel and rifampicin were purchased
EMA guideline also differs from that of the FDA for the fraction ab- from SigmaAldrich (St. Louis, MO), cisapride and rofecoxib were
sorbed (FA) level used to signal high permeability. The EMA sets it purchased from Toronto Research Chemicals (North York, ON,
at >85% whereas FDA uses >90% (FDA, 2000; EMA, 2010). Further it Canada) and lorazepam was purchased from Larodan Fine
has been argued that acids are unfairly assessed with the pH-inter- Chemicals (Malmö, Sweden). For the studies of solubility in FaSSIF,
val used in the solubility definition, since the lower limit of pH 1 SIF powder original was used which was kindly donated by
(reflecting the pH in the stomach) classifies a large number of acids biorelevant.com.
as BCS class 2 although they display high solubility in the small
intestine in vivo which is the compartment where most of the 2.3. Computational prediction of biorelevant solubility: impact of pH
absorption takes place (Dressman et al., 2001; Yazdanian et al., and lipids
2004). This issue has been addressed in the WHO guidance. Phar-
maceutical products containing BCS Class 2 weak acids that are sol- The solubility at pH 1.0 (reflecting stomach) and 6.8 (reflecting
uble at pH 6.8 are according to WHO eligible for the biowaiver small intestine) was predicted with ADMET Predictor (Simula-
procedure, provided that they dissolve rapidly at pH 6.8 and simi- tionsPlus, CA). The potential of lipids to further increase the solu-
larly to the comparator product at pH 1.2 and 4.5 (WHO, 2006). bility in comparison to a pure buffer was forecasted using the
Following the debate on which pH-range to use, more physiologi- solubilization ratio (SR) and two different computational models
cally relevant dissolution media in which the pure buffers are in ADMET Predictor. This procedure was used as we did not have
exchanged to buffers spiked with e.g. bile components to also ad- large enough datasets at hand to validate the accuracy of the differ-
dress solubilization likely to occur in vivo, have been suggested ent solubility models in ADMET Predictor. Therefore, the predicted
226 C.A.S. Bergström et al. / European Journal of Pharmaceutical Sciences 57 (2014) 224–231

Table 1 Table 2
Compounds reported as belonging to only one BCS class in the survey (n = 183). Compounds sorted into two BCS classes in the present literature (n = 59).

Class 1 (n = 70) Class 1/2 (n = 9)

Amlodipinea Glucoseb Proguanilc Amitriptylinea,b
Benznidazolec Imipramineb Promazineb Chlorpromazinea,b
Bisoprolola Ketorolacd Propranololc Digoxinb,c
Buspironeb Labetalole Pyridoxinec Ketoprofena,c
Caffeineb Levodopac Quinaprila Nifedipineb,c
Chloroquinec Levonorgestrelc Quinidineb Piroxicamc,d
Citaloprama Lidocaineb Ramiprila
Valproic acidb,c
Cyclophosphamidec Lithiumc Riboflavinc
Warfarina,b
Desipraminef Lomefloxacinb Rosiglitazoneb
Verapamil HClb
Diazepamc Loratadinea Salbutamolc
Diethylcarbamazinec Meperidineb Salicylic acidb Class 1/3 (n = 20)
Diltiazemb Metoprololf Sertralinea Abacavirb,c Dexamethasoneb Metoclopramideb
Diphenhydramineb Metronidazolec Sildenafila Acetylsalicylic acidb,d Fluconazoleb,e Paracetamolb,f
Disopyramideb Midazolamb Sotalole Atropine sulphatea,b Isoniazidb,e Pravastatinc,e
Donepezila Minocyclineb Stavudinec Captoprilb,c Lamivudineb,e Promethazinea,b
Doxazosina Mirtazapinea Terbinafina Cetirizinec,e Levamisoleb Pyrazinamideb,e
Doxepinb Misoprosolb Theophyllinec,f Chlorpheniramineb Levofloxacinc,e Quinineb
Doxycyclinec,f Nicotinamidec Timolole Clomipheneb Methyldopaa,b
Enalaprila Norethisteronec Tramadola
Ephedrineb Ondansetrona Venlafaxinea Class 1/4 (n = 3)
Ergonovinef Phenobarbitalb,c Zidovudinec Amiloridea,b
Ethinyl estradiolf Phenylalanineb Zolpidema Amoxicillina,b
Ethosuximidec Prednisolonec Acetaminophenc,e
Fluoxetineb Primaquinec,f
Class 2/3 (n = 2)
Class 2 (n = 58) Erythromycina,bc
Aceclofenaca Ibuprofenc Praziquantelc Trimethoprimb,c
Albendazolec Indomethacinf Raloxifeneb
Amiodaroneb Iopanoic acidc Rifampicinc Class 2/4(n = 20)
Atorvastatinb Irbesartana Risperidonea Ciprofloxacina,b Lopinavirb Retinolb
Azithromycinb Itraconazoleb Rofecoxibd Diloxanideb Mebendazoleb Spironolactoneb
Carbamazepinec Ketoconazoleb Simvastatina Efavirenzb Mefloquineb Saquinavirb,c
Carvedilola Lamotriginea Sirolimusb Folic acidb Nelfinavirb,c Sulfadiazineb
Celecoxibg Lansoprazoleb Sulfamethoxazolec Glibenclamideb,g Niclosamideb Sulfasalazineb
Cisaprideb Lorazepama Sulindacd Indinavirb,c Pyrantelb Triclabendazoleb
Clofaziminec Lovastatina Tacrolimusb Ivermectinb Pyrimethamineb
Clopidogrela Mefenamic acidh Talinololb
Class 3/4(n = 5)
Cyclosporineb Meloxicamd Tamoxifenf
Acyclovira,b
Danazolg Montelukastg Terfenadinef
Dapsonec Mycophenolate mofetila Tolmetind Famotidinec,e
Diclofenac sodiumg Nalidixic acidc Furosemideb,c
Diflunisalb Naproxenf Hydrochlorothiazidea,b
Dipyridamoleh Nevirapinec Methotrexatea,b
Ebastinea Nitrofurantoinc a
Bergstrom et al. (2003).
Fenoprofend Ofloxacinb b
Lindenberg et al. (2004).
Flurbiprofenb Oxaprozinb c
Glipizideb Phenazopyridinef Wu and Benet (2005).
d
Griseofulvinc Phenytoinc Yazdanian et al. (2004).
e
Ramirez et al. (2010).
Class 3 (n = 48) f
Butler and Dressman (2010).
Acebutolole b
Dicloxacillin Nadolole
g
Kleberg et al. (2010).
Alendronic acida Didanosinec Neostigminec
Allopurinolc Ergocalciferolc Nifurtimoxc
Anastrozolea Ergometrinec Penicillaminec
Ascorbic acidc Ergotaminec Penicillinsb log DpH6.5 values (distribution of a compound between octanol and
Atenololc,e Fexofenadineb Propylthiouracilc water at pH 6.5) from ADMET Predictor were used to calculate the
Bidisomideb Folinic acidf Pyridostigminec
Biperidenc Gabapentina Ranitidinea
SR at pH 6.5 (the pH of FaSSIF) using the following equation:
Bisphosphonatesb Ganciclovirb Reserpinec
log SR ¼ 0:39 log DpH6:5 þ 3:52 ð1Þ
Cefaclora Hydralazinec Risedronic acida
Cefazolinb Letrozolea Terazosina
The obtained SR was then used to calculate the solubilization capac-
Chloramphenicolc Levetiracetama Tetracyclineb
Cimetidinec Levothyroxinec Thiaminec ity of the buffer and the FaSSIF using the following equation:
Cloxacillinb,c Lisinoprila Topiramatea
Codeinea,c Losartana Valsartanb
SR ¼ SC bs =SC aq ð2Þ
Colchicinec Metforminc Zalcitabineb
where SCbs and SCaq are the solubilization capacity of the bile salt
Class 4 (n = 7)
and water, respectively. Eq. (1) was established by Fagerberg and
Acetazolamidec Cefuroxime axetil a

Aluminium hydroxidec Oxcarbazepinea colleagues in 2010 as a further development of the SR and its rela-
Azathioprinec Ritonavirc tionship to log P presented by Mithani and co-workers in 1996
Cefiximea (Mithani et al., 1996; Fagerberg et al., 2010). In the following, both
a
Ramirez et al. (2010). the FaSSIF solubility and the corresponding solubility in a pure buf-
b
Wu and Benet (2005). fer with the same pH as FaSSIF (pH 6.5) were predicted with ADMET
c
Lindenberg et al. (2004). Predictor. In step 1, the predicted log DpH6.5 values were used to cal-
d
Yazdanian et al. (2004).
e
culate the log SR (Eq. (1)) and the resulting SR was used together
Yang et al. (2007).
f
Bergstrom et al. (2003). with predicted FaSSIF solubility (e.g. buffer pH 6.5 with 0.75 mM
g
Kleberg et al. (2010). phospholipid and 3.0 mM taurocholate) to calculate the concentra-
h
Butler and Dressman (2010). tion possible to achieve in the buffer (Eq. (2)). The calculation was
 C.A.S. Bergström et al. / European Journal of Pharmaceutical Sciences 57 (2014) 224–231 227

Table 3
Physicochemical properties and FA data for BCS class 2 compounds.

Compound Mw log P log D6.5 pKa a/b/n/z/am FA (%)

Albendazole 265.3 3.2 3.2 4.0 b 5a
Tacrolimus 804.0 3.6 3.6 n 15b
Cyclosporine 1202.6 3.0 3.0 n 28a
Amiodarone 645.3 7.2 5.2 8.5 b 30 (F)c*
Lovastatin 404.5 4.5 4.5 n 31b
Dipyridamole 504.6 2.2 1.8 6.7 b 36d
Azithromycin 749.0 3.3 0.5 8.4 b 37a
Clofazimine 473.4 6.8 4.3 9.0 b 45e
Griseofulvin 352.8 2.5 2.5 n 45d
Clopidogrel 321.8 3.4 3.4 4.6 b 50f
Danazol 337.5 3.6 3.6 n 58d
Raloxifene 473.6 5.5 4.1 8.5 b 60g
Carvedilol 406.5 3.9 2.4 8.1 b 65b
Talinolol 363.5 3.5 0.8 9.3 b 65b
Lamotrigine 256.1 2.0 2.0 3.4 b 70h
Risperidone 410.5 2.9 1.4 8.0 b 70g
Simvastatin 418.6 4.9 4.9 n 73g
Ketoconazole 531.4 3.9 3.9 5.7; 4.1 b 75i
Irbesartan 428.5 4.2 4.0 6.7 b 78d
Itraconazole 705.6 5.2 5.2 3.9 b 80j
Carbamazepine 236.3 2.6 2.6 n 85j
Fenoprofen 242.3 3.5 1.4 4.4 a 85k
Lansoprazole 369.4 1.8 1.8 3.2 b 85j
Sulindac 356.4 3.2 0.8 4.1 a 88k
Meloxicam 351.4 2.2 1.1 3.9 a 89k
Diflunisal 250.2 4.0 1.0 3.1 a 90k
Phenazopyridine 213.2 2.8 2.8 5.1 b 90b
Tolmetin 257.3 2.8 0.1 3.8 a 90k
Nitrofurantoin 238.2 0.6 0.3 4.1 a 90(F)a
Phenytoin 252.3 2.2 2.2 8.0 a 90 (F)c
Flurbiprofen 244.3 3.6 1.5 4.3 a 92k
Ibuprofen 206.3 3.7 1.8 4.6 a 92j
Rofecoxib 314.4 2.7 2.7 n 93k
Dapsone 248.3 0.8 0.8 3.2 b 93c
Lorazepam 321.2 2.6 2.6 n 93(F)a
Nevirapine 266.3 1.7 1.7 2.3 b 93 (F)c
Rifampicin 822.9 3.1 3.0 1.4; 7.4 am 93 (F)l
Mycophenolate mofetil 320.3 3.0 1.2 4.4 a 94 (F)c
Oxaprozin 293.3 3.5 1.7 4.6 a 95k
Indomethacin 357.8 3.8 1.5 4.1 a 98k
Naproxen 230.3 3.3 1.3 4.5 a 99k
Cisapride 466.0 3.5 2.6 7.3 b 100j
Atorvastatin 558.6 4.8 3.0 4.7 a 100m
Diclofenac sodium 296.2 4.7 2.2 4.0 a 100n
Glipizide 445.5 2.2 1.3 5.5 a 100b
Ofloxacin 361.4 0.4 0.5 6.1; 7.7 z 100b
Praziquantel 312.4 2.2 2.2 n 100j
Tamoxifen 371.5 6.7 4.8 8.4 b 100b
Sulfamethoxazole 253.3 0.6 0.2 5.8 a 100 (F)c
Aceclofenac 354.2 4.3 1.2 3.2 a N.F.
Celecoxib 381.4 3.5 3.5 n N.F.
Ebastine 469.7 6.7 4.7 8.5 b N.F.
Iopanoic acid 570.9 4.2 2.5 4.8 a N.F.
Mefenamic acid 241.3 5.0 2.9 4.4 a N.F.
Montelukast 586.2 7.3 5.9 5.1 a N.F.
Nalidixic acid 232.2 1.2 0.1 5.5 a N.F.
Sirolimus 914.2 4.5 4.5 n N.F.
Terfenadine 471.7 5.7 3.5 8.8 b N.F.

Abbreviations used: molecular weight (Mw); partition coefficient between octanol and water (log P); partition coefficient between octanol and water at pH 6.5 (log D6.5); acid
(a); base (b); neutral in the pH-range of 2–12 (n); zwitterion (z); ampholyte (am); fraction absorbed (FA); not found (N.F.). *F (bioavailability) low due to low absorption.
a
Perez et al. (2004).
b
Varma et al. (2012).
c
Brunton (2005).
d
Sugano (2011).
e
Mathur et al. (1985).
f
Taubert et al. (2006).
g
Gan et al. (2009).
h
Sanghvi et al. (2003).
i
Zhu et al. (2002).
j
Chu and Yalkowsky (2009).
k
Yazdanian et al. (2004).
l
Loos et al. (1985).
m
Lennernas (2003).
n
Chiou and Barve (1998).
228 C.A.S. Bergström et al. / European Journal of Pharmaceutical Sciences 57 (2014) 224–231

Fig. 1. Computationally predicted pH-dependent dose number for BCS class 2 compounds. The graph presents the logarithm of the dose number (log Do), and hence, log Do of
0 corresponds to Do of 1 (here marked with the dashed line). For compounds with a Do less than this value the maximum dose given is completely soluble. The compounds
are grouped based on their protolytic function with acids, bases, non-ionizable, ampholyte and zwitterion displayed from left to right.

repeated in a second round, where instead of predicted FaSSIF, the and FaSSIF was analysed with Student’s t-test. The obtained solu-
predicted aqueous solubility at pH 6.5 was used in Eq. (2) to calcu- bility data were used to calculate the Do according to Eq. (3).
late the solubility in FaSSIF.
The dose number (Do) was calculated from
3. Results and discussion
Do ¼ ðM 0 =V 0 Þ=C s ð3Þ
Of the 242 BCS compounds extracted from the literature, 183
where M0 is the dose of the compound (here set to the maximum compounds had been classified as belonging to one single class
dose strength), V0 is the volume (here set to 250 mL) and Cs is the (Table 1) and 59 were categorized as belonging to two different
solubility in the medium used (i.e. FaSSIF or buffer pH 6.5) (Oh
et al., 1993). The Do was predicted based on the values obtained
from Eqs. (1) and (2). (a) 2.5 Buffer
FaSSIF
1.5
2.4. Experimental determinations of solubility in FaSSIF and buffer
0.5
logDo

For the twelve non-acidic BCS 2 compounds with a FA > 85% the
apparent solubility was measured in FaSSIF and corresponding -0.5
blank buffer (i.e. FaSSIF without 3 mM taurocholate and 0.75 mM
lecithin) to investigate the solubilization capacity of the mixed li- -1.5
pid aggregates present in the FaSSIF. SIF Powder original was used
-2.5
for the production of FaSSIF according to the protocol provided by
Da cin

flo el

Ca opy ib
e
Ta p a m

le

in

ne

az e

e
Lo prid

in
az on

am idin
ns xife

the manufacturer. For eight of the compounds the apparent solu-

zo

Ne ac

ox
i
n

ep
pi

p
Pr ps

ua
ra

x
ze

ra

en fec
m
o

r
sa

bility was measured using the lDISS Profiler (pION Inc, MA) at
op

iq
m

vi
ra

fa

Ph Ro
Ci

O
Ri

az

37 °C (n = 3), according to a previously published protocol (Fager-

rb
La

berg et al., 2010). The pH of the suspensions (buffer and FaSSIF)

remained constant during the course of the experiment. For loraze- Buffer
(b) 2.5
pam, ofloxacin and rifampicin, a small-scale shake flask method
FaSSIF
was used for the solubility assessment due to low amount of com- 1.5
pound available (lorazepam) or high solubility (ofloxacin and rif-
ampicin). The solubility assay used 300 lL of buffer or FaSSIF and 0.5
logDo

was run in duplicate. Excess amount of solid material was weighed

into glass vials producing a suspension throughout the experiment. -0.5
Pre-heated solvent was added and the vials were put on a plate-
-1.5
shaker (300 rpm) in an incubator (37 °C) and equilibrated for
24 h. The vials were then centrifuged in a temperature controlled
-2.5
Eppendorf centrifuge Model 5043 at 10,000 g, 37 °C for 10 min
ec e
D cin

op ide
La isa b

n
e

m in
flo m

ar vir el
iq le

yr e
e
R son

Ph a m pin
i

ife
op in

and the supernatant was sampled, diluted and determined for

R idin
ox

Ta pic
N ant
pa

az zo
ns pr
xa

az e p

ox
Pr ra

a
ze

ap

m
u

en a z

concentration in a Tecan Sapphire plate reader (Tecan Group,

ra

of

ifa
O

e
Lo

Switzerland). The pH of the suspensions (buffer and FaSSIF)

b
C

remained constant during the course of the experiment. Solubility

data of tamoxifen in buffer and FaSSIF were taken from a previous
Fig. 2. Prediction of solubility in buffer and FaSSIF taking use of the solubilization
study performed in-house (Fagerberg et al., 2010). ratio (SR). (a) Solubility in FaSSIF calculated by Eqs. (1) and (2) using predicted
All solubility data are presented as the mean ± standard devia- buffer solubility from ADMET Predictor, and (b) solubility in buffer calculated by
tion. Statistical difference between the apparent solubility in buffer Eqs. (1) and (2) using predicted solubility in FaSSIF from ADMET Predictor.
 C.A.S. Bergström et al. / European Journal of Pharmaceutical Sciences 57 (2014) 224–231 229

BCS classes in different publications (Table 2). The discrepancy be- possible to extract FA from the literature for 49 of the BCS class
tween publications may be a result of the different guidelines pro- 2 compounds (Table 3). The majority (n = 29) had a reported
vided by EMA and FDA or that a compound may be used in FA > 85%, 11 compounds had a FA of 50–84% and only 9 com-
different maximum doses when indicated for the treatment of dif- pounds had a reported FA < 50%.
ferent diseases. As discussed in the introduction, there are different The protolytic functions of the 58 class 2 compounds (Table 3)
pH-intervals for solubility assessments and FA limits for perme- were equally distributed between acids (37.9%) and bases
ability; however, dose also differs in the two guidelines. The EMA (37.9%), whereas 20.7% were non-ionizable. Only one ampholyte
requires the maximum oral dose administered to be soluble and one zwitterion were included in the dataset. The 29 BCS class
whereas the FDA requires the maximum dose strength. For exam- 2 compounds with FA > 85% were further analysed for their proto-
ple, if the maximum dose strength of a tablet is 100 mg but it is lytic function. A large majority of these were acids (59%), whereas
possible to administer two tablets simultaneously, FDA would base the remaining were composed of bases (21%), non-ionizable com-
the BCS classification on 100 mg whereas EMA would use 200 mg. pounds (14%), and the zwitterion (3%) and ampholyte (3%). This
Therefore, dependent on which of the two guidelines that is used shows that acidic compounds in many cases are classified as BCS
for the classification, differences are likely to appear. A third reason 2 compounds even though they are absorbed to a high extent
for the differences found in the literature may be that different (>85%) indicating a considerable risk for incorrect classification of
methods have been used when investigating solubility and acids as being compounds with solubility-limited absorption.
permeability. Predicted buffer solubility clearly identified the impact of pH on
The 183 compounds designated to a single BCS class were dis- solubility (Fig. 1). For fifteen of the acidic BCS class 2 compounds a
tributed as follows: 38.2% class 1 compounds, 31.7% class 2 com- Do < 1 was predicted at pH 6.8 compared to a Do > 1 at pH 1.0. It
pounds, 26.2% class 3 compounds and 3.8% class 4 compounds should be noted that all BCS class 2 acids had FA > 85% (Table 3);
(Table 1). For the remainder of this work the analysis focused on hence, in vivo these acids are not likely to display solubility-limited
the compounds sorted as BCS class 2. These are the compounds absorption. As expected the pH dependency for bases resulted in
for which in vivo conditions (by virtue of the pH gradient and/or lower Do at low pH. For many of the BCS class 2 bases Do was pre-
the naturally available lipids in the intestine) can improve solubil- dicted to be less than 1 at pH 1 but showed Do > 1 at pH 6.8 (Fig. 1).
ity to the extent that the complete dose becomes soluble. It was This trend may be useful to identify bases that risk precipitating

Fig. 3. Experimentally determined solubility and dose number for non-acidic compounds with FA > 85%. (a) Solubility data, here presented as the mean ± standard deviation.
The lipophilicity of the compounds (log D6.5) increases to the right. p < 0.05 (*), p < 0.01 (**), p < 0.001 (***). (b) Of the 12 compounds investigated, 4 compounds had a Do < 1 in
buffer and/or FaSSIF. When the DCS volume criterion was used (i.e. 500 mL instead of the BCS volume of 250 mL) also phenazopyridine and dapsone obtained Do < 1 in FaSSIF,
but not in the buffer. (c) Experimentally determined Do (grey), Do calculated based on ADMET Predictor buffer model (white) and buffer Do calculated based on solubilization
ratio (Eqs. (1) and (2)) (black). (d) Experimentally determined Do (grey), Do calculated based on ADMET Predictor FaSSIF model (white) and FaSSIF Do calculated based on
solubilization ratio (Eqs. (1) and (2)) (black).
230 C.A.S. Bergström et al. / European Journal of Pharmaceutical Sciences 57 (2014) 224–231

when the gastric content empties into the higher pH found in the as the pharmaceutical industry, shown by e.g. that biowaivers
duodenum and jejunum. Indeed, several of these (e.g. albendazole, are not accepted by all authorities and by the differences in the
itraconazole and ketoconazole) have been shown to precipitate at guidelines by EMA, FDA and WHO. The computational and experi-
least partially in the small intestine (Jung et al., 1998; Mellaerts mental analyses of BCS class 2 compounds identified that acidic
et al., 2008; Psachoulias et al., 2011). The neutral compounds were, drug molecules are at risk for being deemed having solubility-lim-
as expected, unaffected by the pH. For these compounds, it may be ited absorption by the current BCS classification although complete
more relevant to investigate the impact of the mixed micelles se- absorption is found in vivo. This further supports the WHO guid-
creted by bile on their solubility (Fagerberg et al., 2010; Fagerberg ance in which pharmaceutical products are eligible for biowaivers
et al., 2012). In particular this is important for highly lipophilic if the API is a weak acid for which the oral dose is soluble at pH 6.8
compounds. Lipids have been suggested to significantly impact and the dissolution profile (pH1.2, 4.5 and 6.8) is similar to the
the solubility of drug molecules with a log P > 3 or a log D6.5 > 2 comparator product. In the light of our findings, this combination
(Dressman and Reppas, 2000; Bergstrom et al., 2007; Fagerberg of a solubility measurement at pH 6.8 for weak acidic BCS class 2
et al., 2010; Gamsiz et al., 2010; Ottaviani et al., 2010). We there- compounds and a formulation dissolution performance in the pH
fore analysed whether computational tools predicted lipids to be interval of the gastrointestinal tract is an attractive solubility/dis-
important for the solubility of the remaining 12 non-acidic BCS solution profiling approach to eligible products for biowaivers. Fur-
class 2 compounds. It was found that these models predicted most ther, lipids present in the intestine may increase the solubilization
of them to have increased solubility in FaSSIF as compared to the of compounds in the intestinal fluid and herein it was found that
corresponding buffer (Fig. 2). The Do of tamoxifen (Fig. 2a, predic- compounds having log DpH6.5 P 3.0 had apparent solubility in FaS-
tion based on Eqs. (1) and (2) using buffer solubility from ADMET SIF allowing the complete oral dose to become soluble. However, a
Predictor) and cisapride (Fig. 2b, prediction based on Eqs. (1) and larger number of non-acidic poorly soluble compounds need to be
(2) using FaSSIF solubility from ADMET Predictor) were predicted studied to conclude on possible modifications of the BCS solubility
to decrease to less than 1 in simulated intestinal fluids. This com- definition for such compounds. In conclusion, a more relevant pH
putational exercise indicated that lipids were likely to improve the restriction for acids (i.e. assessing solubility solely at the pH of
solubility of a number of the compounds, maybe even to the extent the small intestinal fluid) and/or physiologically relevant medium
allowing the oral dose to become completely dissolved. Further- with lipids present, such as FaSSIF, identified that 20 of the 29 BCS
more, three of the 12 compounds had log P > 3 and eight of them class 2 compounds with FA > 85% were expected to be completely
displayed log D6.5 > 2, and hence, also the lipophilicity profile of dissolved in human intestinal fluid. From the analyses presented
the compounds pointed at the likelihood of significant solubiliza- herein, global harmonisation of guidelines and the implementation
tion in the lipid aggregates naturally present in the intestinal fluid. of more physiologically relevant media used in the in vitro assess-
We therefore experimentally determined the solubility in buffer ments of solubility and dissolution testing are suggested as two
and FaSSIF of these compounds and found that all of the com- important steps towards an increased number of designated
pounds obtained equal or higher dissolution rate and apparent sol- biowaivers.
ubility in FaSSIF as compared to blank buffer (Fig. 3a). As expected,
the solubility in FaSSIF was equal to that observed in the buffer for Disclaimer
ofloxacin, which was the most hydrophilic compound of the 12
studied with a log DpH6.5 of -0.5. For the most lipophilic compound, The views and conclusions presented in this paper represent
tamoxifen, the solubility in FaSSIF was 26-fold higher than in the those of the authors and not necessarily those of the Swedish Med-
corresponding buffer. The dose number was then investigated as ical Product Agency where two of the authors are employed.
a mean to link dose to solubility (Fig. 3b). Indeed, for the com-
pounds explored here, four of the 12 compounds obtained a
Acknowledgements
Do < 1 in FaSSIF. Two of these, lorazepam and ofloxacin, were also
completely dissolved in the buffer, and hence, the experimental
Financial support from the Swedish Research Council (Grant
determinations to a large extent confirmed the computational pre-
621-2008-3777) and The Swedish Agency for Innovation Systems
dictions (Fig. 3c and d). The other two compounds, rifampicin and
(Grant 2010-00966) is highly appreciated. We are grateful to
tamoxifen, were the most lipophilic compounds measured in this
biorelevant.com for providing us the SIF original powder used for
study (log D6.5 P 3.0), confirming the importance of having solubi-
the solubility experiments, and Simulations Plus (Lancaster, CA)
lizing aggregates present to better mimic in vivo dissolution of
for providing us with a reference site license for the software
highly lipophilic compounds. All other compounds received a high-
ADMET Predictor.
er solubility in FaSSIF than in buffer, although it was not high en-
ough to produce a Do below 1. To summarize, the computational
References
and experimental efforts employed herein, which used more phys-
iologically relevant conditions to investigate the dose possible to Amidon, G.L., Lennernas, H., Shah, V.P., Crison, J.R., 1995. A theoretical basis for a
dissolve for BCS class 2 compounds, indicated complete dissolution biopharmaceutic drug classification: the correlation of in vitro drug product
in intestinal fluid for 20 of the 29 BCS class 2 compounds with dissolution and in vivo bioavailability. Pharm. Res. 12, 413–420.
Benet, L.Z., 2013. The role of BCS (biopharmaceutics classification system) and
FA > 85%. When the DCS solubility criterion (i.e. 500 mL instead BDDCS (Biopharmaceutics Drug Disposition Classification System) in drug
of 250 mL of solvent) was used additional two compounds (dap- development. J. Pharm. Sci. 102, 34–42.
sone and phenazopyridine) received Do < 1. Bergstrom, C.A., Strafford, M., Lazorova, L., Avdeef, A., Luthman, K., Artursson, P.,
2003. Absorption classification of oral drugs based on molecular surface
properties. J. Med. Chem. 46, 558–570.
Bergstrom, C.A., Wassvik, C.M., Johansson, K., Hubatsch, I., 2007. Poorly soluble
4. Conclusion marketed drugs display solvation limited solubility. J. Med. Chem. 50, 5858–
5862.
Brunton, L.L., 2005. Goodman and Gilman’s The Pharmacological Basis of
In this study we identified that the BCS is not used to its full po- Therapeutics: Digital Edition. McGraw-Hill Professional Publishing.
tential. The number of biowaivers filed for and designated is still, Butler, J.M., Dressman, J.B., 2010. The Developability Classification System:
more than 10 years after the introduction of the regulatory guide- application of biopharmaceutics concepts to formulation development. J.
Pharm. Sci. 99, 4940–4954.
lines, relatively low. We speculate that one reason for this is that Chiou, W.L., Barve, A., 1998. Linear correlation of the fraction of oral dose absorbed
the regulatory authorities are not globalized to the same extent of 64 drugs between humans and rats. Pharm. Res. 15, 1792–1795.
 C.A.S. Bergström et al. / European Journal of Pharmaceutical Sciences 57 (2014) 224–231 231

Chu, K.A., Yalkowsky, S.H., 2009. An interesting relationship between drug Ottaviani, G., Gosling, D.J., Patissier, C., Rodde, S., Zhou, L., Faller, B., 2010. What is
absorption and melting point. Int. J. Pharm. 373, 24–40. modulating solubility in simulated intestinal fluids? Eur. J. Pharm. Sci. 41, 452–
Clarysse, S., Brouwers, J., Tack, J., Annaert, P., Augustijns, P., 2011. Intestinal drug 457.
solubility estimation based on simulated intestinal fluids: comparison with Perez, M.A., Sanz, M.B., Torres, L.R., Avalos, R.G., Gonzalez, M.P., Diaz, H.G., 2004. A
solubility in human intestinal fluids. Eur. J. Pharm. Sci. 43, 260–269. topological sub-structural approach for predicting human intestinal absorption
Cook, J.A., Bockbrader, H., 2002. An industrial implementation of the of drugs. Eur. J. Med. Chem. 39, 905–916.
biopharmaceutics classification system. Dissol. Technol. 9, 6–9. Polli, J.E., Abrahamsson, B.S., Yu, L.X., Amidon, G.L., Baldoni, J.M., Cook, J.A., Fackler,
Cook, J.A., Davit, B.M., Polli, J.E., 2010. Impact of biopharmaceutics classification P., Hartauer, K., Johnston, G., Krill, S.L., Lipper, R.A., Malick, W.A., Shah, V.P., Sun,
system-based biowaivers. Mol. Pharm. 7, 1539–1544. D., Winkle, H.N., Wu, Y., Zhang, H., 2008. Summary workshop report:
Davit, B.M., Chen, M.L., Conner, D.P., Haidar, S.H., Kim, S., Lee, C.H., Lionberger, R.A., bioequivalence, biopharmaceutics classification system, and beyond. AAPS J.
Makhlouf, F.T., Nwakama, P.E., Patel, D.T., Schuirmann, D.J., Yu, L.X., 2012. 10, 373–379.
Implementation of a reference-scaled average bioequivalence approach for Polli, J.E., Yu, L.X., Cook, J.A., Amidon, G.L., Borchardt, R.T., Burnside, B.A., Burton, P.S.,
highly variable generic drug products by the US Food and Drug Administration. Chen, M.L., Conner, D.P., Faustino, P.J., Hawi, A.A., Hussain, A.S., Joshi, H.N., Kwei,
AAPS J. 14, 915–924. G., Lee, V.H., Lesko, L.J., Lipper, R.A., Loper, A.E., Nerurkar, S.G., Polli, J.W.,
Dressman, J., Butler, J., Hempenstall, J., Reppas, C., 2001. The BCS: where do we go Sanvordeker, D.R., Taneja, R., Uppoor, R.S., Vattikonda, C.S., Wilding, I., Zhang, G.,
from here? Pharm. Tech. 25, 68–76. 2004. Summary workshop report: biopharmaceutics classification system–
Dressman, J.B., Reppas, C., 2000. In vitro-in vivo correlations for lipophilic, poorly implementation challenges and extension opportunities. J. Pharm. Sci. 93,
water-soluble drugs. Eur. J. Pharm. Sci. 11 (Suppl 2), S73–80. 1375–1381.
EMA, 2010. Committe for Medicinal Products for Human Use. Guideline on the Pouton, C.W., 2006. Formulation of poorly water-soluble drugs for oral
investigation of bioequivalence. <http://www.emea.europa.eu/docs/en_GB/ administration: physicochemical and physiological issues and the lipid
document_library/Scientific_guideline/2010/01/WC500070039.pdf> (accessed formulation classification system. Eur. J. Pharm. Sci. 29, 278–287.
3.05.13). Psachoulias, D., Vertzoni, M., Goumas, K., Kalioras, V., Beato, S., Butler, J., Reppas, C.,
Fagerberg, J.H., Al-Tikriti, Y., Ragnarsson, G., Bergstrom, C.A., 2012. Ethanol effects 2011. Precipitation in and supersaturation of contents of the upper small
on apparent solubility of poorly soluble drugs in simulated intestinal fluid. Mol. intestine after administration of two weak bases to fasted adults. Pharm. Res.
Pharm. 9, 1942–1952. 28, 3145–3158.
Fagerberg, J.H., Tsinman, O., Sun, N., Tsinman, K., Avdeef, A., Bergstrom, C.A., 2010. Ramirez, E., Laosa, O., Guerra, P., Duque, B., Mosquera, B., Borobia, A.M., Lei, S.H.,
Dissolution rate and apparent solubility of poorly soluble drugs in biorelevant Carcas, A.J., Frias, J., 2010. Acceptability and characteristics of 124 human
dissolution media. Mol. Pharm. 7, 1419–1430. bioequivalence studies with active substances classified according to the
FDA, 2000. US Food and Drug Administration. Guidance for industry: waiver of Biopharmaceutic Classification System. Br. J. Clin. Pharmacol. 70, 694–702.
in vivo bioavailability and bioequivalence studies for immediate-release solid Rinaki, E., Dokoumetzidis, A., Valsami, G., Macheras, P., 2004. Identification of
oral dosage forms based on a biopharmaceutics classification system. <http:// biowaivers among Class II drugs: theoretical justification and practical
www.fda.gov/downloads/Drugs/.../Guidances/ucm070246.pdf> (accessed examples. Pharm. Res. 21, 1567–1572.
3.05.13). Rinaki, E., Valsami, G., Macheras, P., 2003. Quantitative biopharmaceutics
Gamsiz, E.D., Ashtikar, M., Crison, J., Woltosz, W., Bolger, M.B., Carrier, R.L., 2010. classification system: the central role of dose/solubility ratio. Pharm. Res. 20,
Predicting the effect of fed-state intestinal contents on drug dissolution. Pharm. 1917–1925.
Res. 27, 2646–2656. Sanghvi, T., Ni, N., Mayersohn, M., Yalkowsky, S.H., 2003. Predicting passive
Gan, J., Ruan, Q., He, B., Zhu, M., Shyu, W.C., Humphreys, W.G., 2009. In vitro intestinal absorption using a single parameter. Qsar. Comb. Sci. 22, 247–257.
screening of 50 highly prescribed drugs for thiol adduct formation–comparison Soderlind, E., Karlsson, E., Carlsson, A., Kong, R., Lenz, A., Lindborg, S., Sheng, J.J.,
of potential for drug-induced toxicity and extent of adduct formation. Chem. 2010. Simulating fasted human intestinal fluids: understanding the roles of
Res. Toxicol. 22, 690–698. lecithin and bile acids. Mol. Pharm. 7, 1498–1507.
Jung, H., Medina, L., Garcia, L., Fuentes, I., Moreno-Esparza, R., 1998. Absorption Sugano, K., 2011. Fraction of a dose absorbed estimation for structurally diverse low
studies of albendazole and some physicochemical properties of the drug and its solubility compounds. Int. J. Pharm. 405, 79–89.
metabolite albendazole sulphoxide. J. Pharm. Pharm. 50, 43–48. Takagi, T., Ramachandran, C., Bermejo, M., Yamashita, S., Yu, L.X., Amidon, G.L., 2006. A
Kleberg, K., Jacobsen, J., Mullertz, A., 2010. Characterising the behaviour of poorly provisional biopharmaceutical classification of the top 200 oral drug products in
water soluble drugs in the intestine: application of biorelevant media for the United States, Great Britain, Spain, and Japan. Mol. Pharm. 3, 631–643.
solubility, dissolution and transport studies. J. Pharm. Pharm. 62, 1656–1668. Taubert, D., von Beckerath, N., Grimberg, G., Lazar, A., Jung, N., Goeser, T., Kastrati,
Lennernas, H., 2003. Clinical pharmacokinetics of atorvastatin. Clin Pharmacokinet A., Schomig, A., Schomig, E., 2006. Impact of P-glycoprotein on clopidogrel
42, 1141–1160. absorption. Clin. Pharmacol. Ther. 80, 486–501.
Lennernas, H., Abrahamsson, B., 2005. The use of biopharmaceutic classification of Varma, M.V., Gardner, I., Steyn, S.J., Nkansah, P., Rotter, C.J., Whitney-Pickett, C., Zhang,
drugs in drug discovery and development: current status and future extension. H., Di, L., Cram, M., Fenner, K.S., El-Kattan, A.F., 2012. pH-Dependent solubility and
J. Pharm. Pharm. 57, 273–285. permeability criteria for provisional biopharmaceutics classification (BCS and
Lindenberg, M., Kopp, S., Dressman, J.B., 2004. Classification of orally administered BDDCS) in early drug discovery. Mol. Pharm. 9, 1199–1212.
drugs on the World Health Organization Model list of Essential Medicines WHO, 2006. Proposal to waive in vivo bioequivalence requirements for the WHO
according to the biopharmaceutics classification system. Eur. J. Pharm. model list of essential medicines immediate release, solid oral dosage forms. In:
Biopharm. 58, 265–278. Fortieth report of the WHO Expert Committee on Specifications for
Loos, U., Musch, E., Jensen, J.C., Mikus, G., Schwabe, H.K., Eichelbaum, M., 1985. Pharmaceutical Preparations. Geneva, World Health Organization. Technical
Pharmacokinetics of oral and intravenous rifampicin during chronic Report Series, no. 937, Annex 8.
administration. Klinische Wochenschrift 63, 1205–1211. Wu, C.Y., Benet, L.Z., 2005. Predicting drug disposition via application of BCS:
Mathur, A., Venkatesan, K., Bharadwaj, V.P., Ramu, G., 1985. Evaluation of transport/absorption/elimination interplay and development of a
effectiveness of clofazimine therapy. I. Monitoring of absorption of biopharmaceutics drug disposition classification system. Pharm. Res. 22, 11–23.
clofazimine from gastrointestinal tract. Indian J. Lepr. 57, 146–148. Yang, Y., Faustino, P.J., Volpe, D.A., Ellison, C.D., Lyon, R.C., Yu, L.X., 2007.
Mellaerts, R., Mols, R., Kayaert, P., Annaert, P., Van Humbeeck, J., Van den Mooter, G., Biopharmaceutics classification of selected beta-blockers: solubility and
Martens, J.A., Augustijns, P., 2008. Ordered mesoporous silica induces pH- permeability class membership. Mol. Pharm. 4, 608–614.
independent supersaturation of the basic low solubility compound itraconazole Yazdanian, M., Briggs, K., Jankovsky, C., Hawi, A., 2004. The ‘‘high solubility’’
resulting in enhanced transepithelial transport. Int. J. Pharm. 357, 169–179. definition of the current FDA Guidance on Biopharmaceutical Classification
Mithani, S.D., Bakatselou, V., TenHoor, C.N., Dressman, J.B., 1996. Estimation of the System may be too strict for acidic drugs. Pharm. Res. 21, 293–299.
increase in solubility of drugs as a function of bile salt concentration. Pharm. Yu, L.X., Amidon, G.L., Polli, J.E., Zhao, H., Mehta, M.U., Conner, D.P., Shah, V.P., Lesko,
Res. 13, 163–167. L.J., Chen, M.L., Lee, V.H., Hussain, A.S., 2002. Biopharmaceutics classification
Nicolaides, E., Symillides, M., Dressman, J.B., Reppas, C., 2001. Biorelevant system: the scientific basis for biowaiver extensions. Pharm. Res. 19, 921–925.
dissolution testing to predict the plasma profile of lipophilic drugs after oral Zhu, C., Jiang, L., Chen, T.M., Hwang, K.K., 2002. A comparative study of artificial
administration. Pharm. Res. 18, 380–388. membrane permeability assay for high throughput profiling of drug absorption
Oh, D.M., Curl, R.L., Amidon, G.L., 1993. Estimating the fraction dose absorbed from potential. Eur. J. Med. Chem. 37, 399–407.
suspensions of poorly soluble compounds in humans: a mathematical model.
Pharm. Res. 10, 264–270.