Sei sulla pagina 1di 15

The Developability Classification System: Application of

Biopharmaceutics Concepts to Formulation Development


JAMES M. BUTLER,1 JENNIFER B. DRESSMAN2
1

Pharmaceutical Development, GlaxoSmithKline R&D, Harlow, UK

Institute of Pharmaceutical Technology, Goethe University, Frankfurt/Main, Germany

Received 21 December 2009; revised 9 March 2010; accepted 16 April 2010


Published online 2 June 2010 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jps.22217
ABSTRACT: A revised classification system for oral drugs was developed using the biopharmaceutics classification system (BCS) as a starting point. The revised system is designed to have
a greater focus on drug developability. Intestinal solubility, the compensatory nature of
solubility and permeability in the small intestine and an estimate of the particle size needed
to overcome dissolution rate limited absorption were all considered in the revised system. The
system was then validated by comparison with literature on the in vivo performance of a number
of test compounds. Observations on the test compounds were consistent with the revised
classification, termed the developability classification system (DCS), showing it to be of greater
value in predicting what factors are critical to in vivo performance than the widely used BCS.
2010 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 99:49404954, 2010

Keywords: biopharmaceutics classification system (BCS); bioequivalence; oral absorption;


developability; biorelevant media; solubility; permeability; dissolution rate

INTRODUCTION
The introduction of the biopharmaceutics classification system (BCS) in the 1990s made a significant
impact on the development of immediate release (IR)
oral dosage forms, enabling the use of in vitro data
rather than in vivo human studies for establishing
the bioequivalence of low risk (BCS class I) compounds.1,2
The BCS also provides a framework to consider key
factors (dose, solubility, permeability and dissolution
rate) that may influence in vivo performance. The
significance of these factors is broader than identifying drug products suitable for biowaivers, as they by
inference also define the critical quality attributes
(CQAs) which determine in vivo performance. These
need to be understood when optimising an oral drug
through the product development process, particularly in the context of quality by design (QbD).3
Due to the significant regulatory element to the
BCS, the classification system rightly adopts a
relatively conservative approach, particularly in the
assessment of when solubility and/or dissolution rate
(which are strongly dependent on product attributes)
are critical in limiting oral absorption. This is because
Correspondence to: James M. Butler (Telephone: 44-1279-631491; Fax: 44-1279-64-3953; E-mail: james.m.butler@gsk.com)
Journal of Pharmaceutical Sciences, Vol. 99, 49404954 (2010)
2010 Wiley-Liss, Inc. and the American Pharmacists Association

4940

in the context of protecting patients, it is more


important to ensure changes to products that result in
altered in vivo performance are correctly classified
rather than the misclassification of those that do not.
Changes to a drug product generally have less
influence over permeability as this is a property of the
drug molecule that is less likely to alter with product
and process change. This makes it easier to identify
a theoretical high/low permeability boundary (e.g.
permeability equivalent to 90% fraction absorbed)
that is meaningful in determining if permeability is
partly rate limiting to oral absorption, and also makes
permeability less likely to be related to the CQAs of a
drug product.
Exceptionally, excipients may have an impact on
in vivo permeability, either directly on the active or
passive transit of the drug across the gut wall or
indirectly by altering the GI transit/residence time.
Although there are known examples where this
occurs,4,5 these are relatively rare, and tend to
require a large amount of specific high risk excipients
for any significant in vivo change.
Extensions to the biowaiver classes of the BCS
have been proposed.6 These include the extension of
biowaivers to class III compoundsprovided changes
to excipients are unlikely to alter drug permeability,7
and to some class II weak acids on the basis that these
drugs often have adequate solubility and permeability in the upper small intestine, so their failure to
meet the high solubility criteria at gastric pH has

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 12, DECEMBER 2010

THE DEVELOPABILITY CLASSIFICATION SYSTEM

little impact in vivo.8 These extensions have largely


been adopted by recent WHO guidance.9
In order to ease the identification of compounds
as high or low permeability, Wu and Benet10 have
proposed metabolic clearance as an appropriate
surrogate for permeability, and have highlighted
the considerable overlap between the cut-off for low/
high permeability and the tendency of drugs to be
cleared unchanged (if low permeability) or via
metabolic pathways (if high permeability). The
reason for this overlap is postulated to be the
permeability-dependent ability of a drug to access
metabolic enzymes within cells.
In this paper a revised version of the BCS is
presented, designed not so much as a regulatory
classification for the assurance of bio-equivalence, but
one that will more accurately categorise drugs
according to the factors limiting their oral absorption.
This provides a more appropriate classification
system for the issues relating to oral product
development in the framework of Quality by Design
(QbD). Despite the change of focus from the BCS, such
a system may still provide a scientific framework
for discussion with regulators as to the risk of
bio-inequivalence for changes to drugs that currently
fall outside the current BCS I category, particularly
those in BCS class II, which is the most frequent class
for orally active new chemical entities (NCEs).

THEORETICAL/BACKGROUND
Modelling oral absorption has been the focus of
several key papers in the last decade.
One simple concept, widely used in the early stages
of drug development, is that of maximum absorbable
dose (MAD). Several versions of MAD exist, using
different assumptions and means of estimating
permeability. In one version proposed by Curatolo11
the derivation uses an absorption rate constant (KA),
whilst another by Sun et al.12 uses an estimate of the
effective human jejunal permeability ( Peff).
MAD S  KA  V  T

(1)

MAD Peff; human  S  A  Tsi

(2)

where S is the solubility, V the fluid volume (250 mL),


Tsi the transit time for the absorption site (3.32 h for
the small intestine) and A is the absorption surface
area (7.54  104 cm2).
One of the implications of the MAD equation is
that permeability and solubility are compensatory;
that is, high permeability may offset low solubility
in determining the maximum dose above which
solubility in the GI tract becomes restrictive to
absorption. Further evidence that this may be a
useful assumption, at least for high permeability
DOI 10.1002/jps

4941

drugs as demonstrated by Fagerholm et al.13 is that


the permeability of deuterium in the human jejunum
is at a rate close to the high/low boundary in the BCS.
Assuming deuterium has a similar intestinal permeability to water, this means that high permeability
drugs are generally absorbed quicker than the fluid
from the small intestine, thus raising the dose at
which solubility becomes limiting, whereas low
permeability drugs tend to be absorbed at a slower
rate than the water in the small intestine. In this
context, Eq. (2) incorporates a particularly useful
measure of permeability, as it is derived from the
measurement of the rate of disappearance of drug
from the jejunum.
Another model that is widely used in oral absorption modelling generates three dimensionless numbers Do, Dn and An for assessment of whether dose/
solubility ratio, dissolution rate and/or permeability
are likely to limit oral absorption in the GI tract.14,15
In this model, several important concepts are
expressed:
(1) A drug will be dissolution rate limited if the
rate of dissolution is too slow for all of the drug
particles to dissolve during the time for transit
past the absorption site. For simplicity, the
model used a small intestinal transit time to
estimate residence time at the absorption site.
Particle size and intestinal solubility are the
two key product dependent variables.
(2) A drug will be solubility limited if there is
inadequate fluid within the GI tract to dissolve
the administered dose. For simplicity, the
model used the minimum aqueous solubility
in the physiological pH range (1.27.5), and a
fluid volume of 250 mL, matching that of the
BCS. Dose and solubility are the two key product dependent variables.
(3) A drug will be permeability limited if there is an
inadequate rate of transfer from the gut lumen,
across the gut wall. Permeability is the key
drug dependent variable.
By combining these dimensionless numbers it is
possible to estimate the fraction absorbed from the GI
tract. The estimate of fraction absorbed also incorporates the concept from the MAD equation, that
permeability and solubility are compensatory.
Such concepts for modelling oral absorption
have also been used in combination with a compartmental model of the gastro-intestinal tract to form
some of the logic behind the commercially available
modelling tool Gastro-Plus.16 Other models for oral
absorption are also available, for instance based on
modelling gastro-intestinal fluid flow, which forms
the basis of the alternative commercial model
PK-Sim.17
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 12, DECEMBER 2010

4942

BUTLER AND DRESSMAN

Optimal Use of Solubility as an Input Into Oral


Absorption Models
The BCS uses a definition of solubility based on the
lowest solubility in 250 mL in the physiological pH
range. This is likely to be an underestimate of the
actual solubility experienced in vivo as many drugs
have pH dependent solubility, and even for a drug
with pH independent solubility in the physiologically
relevant range, solubilisers present in the gut from
gastro-intestinal secretions and food intake contribute to drug solubilisation.
To improve the prediction of GI dissolution, Dressman and coworkers18,19 have extensively promoted
the concept of biorelevant dissolution media containing solubilisers such as bile acids for the assessment
of drug developability and formulation performance
of poorly soluble drugs. These include fasted state
simulated intestinal fluid (FaSSIF), fed state simulated intestinal fluid (FeSSIF), fasted state simulated
gastric fluid (FaSGF)20 and several fed state simulated gastric fluids (FeSSGF).21
As well as making dissolution tests more biorelevant, these media also make it possible to improve the
estimate of solubility used in oral absorption models.
For example, work by Charman and coworker22 used
FaSSIF and FeSSIF solubility and the Do, Dn, An,
model described above to explain the in vivo performance of the poorly soluble drug halofantrine in the
fasted and fed state. The sizable food effect could be
related to the improved solubility and dissolution
rate under fed state conditions.
The possibility of using biorelevant media rather
than buffers and more relevant volumes than the
250 mL in the BCS or derived versions of the BCS has
already been discussed in a previous paper.23
Poorly soluble drugs often have greater oral
absorption in the fed state due to higher fluid volumes
and greater potential for solubilisation. Therefore, if
complete absorption can be achieved fasted, solubility
related food effects can be eliminated. As a result, the
use of an estimate of intestinal solubility in the fasted
state is particularly significant when predicting the
extent of absorption (AUC), as it is generally desirable
for patient convenience to be able to administer a drug
either with or without food and get equivalent
pharmacokinetic and therapeutic responses.
Obtaining an estimate of jejunal solubility is
important as it is in the upper small intestine that
the fluid volumes for drug dissolution reach a
maximum and it is where most immediate release
drugs are absorbed. With the notable exception of
weak bases (where an estimate of gastric solubility
will also be highly significant), upper small intestinal
solubility is more significant than gastric solubility in
determining whether a drug has adequate solubility
and dissolution rate for complete extent of absorption.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 12, DECEMBER 2010

Gastric solubility (and dissolution rate) may still be of


significance to rate of in vivo absorption (Cmax, Tmax)
for low solubility drugs.
In order to get an even better estimate of in vivo
solubility, there have been a several proposals to
further modify the media proposed by Dressman and
Reppas, including the use of more complex mixtures
of bile salts,24 addition of components to mimic lipids
from food intake,25 adoption of a more in vivo relevant
intestinal buffer26 and the use of naturally occurring
surface active agents in gastric media.27 Although an
estimate of intestinal solubility can be made from the
use of these artificial media to simulate GI conditions,
the use of aspirates extracted from human or animal
gastro-intestinal tract as a media for obtaining an
estimate of the true gastro-intestinal solubility has
also been advocated.28,29 A review of methods to
estimate solubility in the gastro-intestinal tract
covering many of these considerations, is also
available.30
Optimal Use of Permeability as an Input Into Oral
Absorption Models
Permeability may be assessed by a variety of
techniques during the pharmaceutical development
of a new drug. Although it is possible and reasonable
to use a cut off between high and low permeability in
terms of fraction absorbed as defined by the BCS, a
challenge is to find a suitable permeability measurement which has an adequate correlation to fraction
absorbed, but is also simple to perform. Methods
reported in the literature include those based on
in silico drug properties such as log D and hydrogen
bonding potential,31,32 passive diffusion across an
artificial membrane,33 in vitro permeability across
cell lines,3436 across excised human or animal
tissue37,38 and permeability as determined by
in situ perfusion techniques.39 Simpler methods are
preferred during the early phases of product development and for assessment of drug candidates,
whereas in later development more complex measurements may be performed. For formal BCS
classification, there is specific guidance on what is
an acceptable method for assessing permeability,
both in terms of the method used and method
validation.2
In early development, higher throughput, simpler
to use methods, not generally recognised in the BCS
classification guidance, are of particular value, both
as a qualitative (e.g. into high/medium/low bins) and
as a quantitative (permeability expressed as a
number) measure. The usefulness of any quantitative
value can further be enhanced if it can be correlated to
a meaningful measure of in situ human permeability
or fraction absorbed.
A limited data set of human permeability determined via a jejunal perfusion technique is available in
DOI 10.1002/jps

THE DEVELOPABILITY CLASSIFICATION SYSTEM

the literature.40 This is particularly valuable as it


provides a means to correlate any permeability
measurement to a meaningful permeability for both
poorly and well absorbed drugs, whereas fraction
absorbed data is only discriminatory for poorly
absorbed (<90% fraction absorbed) compounds. By
correlating other permeability data (however determined) to the human jejunal permeability data set,
it is possible to obtain a qualitative estimate of
permeability, not just for low permeability compounds, but for high permeability compounds as
well. Examples of this kind of work in the literature
include the correlation of log D-based in silico models,41,42 artificial membranes such as PAMPA,43
in vitro cell lines44 and in vivo rat perfusion4547 to
the human jejunal permeability data set.

METHOD: DEVISING THE MODIFIED


CLASSIFICATION SYSTEM
The modified classification aimed to incorporate the
following concepts:
(1) An estimate of human fasted intestinal solubility (e.g. by using FaSSIF) as the primary
measure of in vivo solubility useful for the
prediction of the extent of human absorption.
(2) A solubility limited absorbable dose (SLAD)
concept, based on the idea that for class II drugs
at least, permeability and solubility are compensatory.
(3) Dissolution rate, expressed as a target drug
particle size rather than dose/solubility ratio,
provides a better means of assessing the development risks and CQAs for drugs with dissolution rate limited extent of absorption.
The incorporation of concepts (1) and (2) is
illustrated in Figure 1. Note the significant deviation

Figure 1. Modifying the BCS for more realistic volumes


of fluid available in the GI tract and the compensatory
nature of permeability on low solubility (modifications from
the BCS to DCS are shown in blue).
DOI 10.1002/jps

4943

from the BCS with the addition of the two sub classes
for class II and the emphasis of the revised system
on the prediction of extent, rather than rate of oral
absorption.
The solubility limited absorbable dose is represented by the boundary between class IIa and IIb for
high permeability drugs, and the boundary between
class III and IV for low permeability drugs. It can be
expressed as:
SLAD Ssi  V  Mp

(3)

where Ssi is the estimate of small intestine solubility,


V the fluid volume (500 mL) and Mp is the permeability dependent multiplier. For a high permeability
drug, Mp is equal to the absorption number (An), but
it is kept at unity for low permeability drugs.
The measure of solubility would typically be that of
saturation solubility in fasted state simulated intestinal fluids, although other measures of intestinal
solubility such as the use of aspirates taken via
intubation of human volunteers or from animals may
be appropriate.
For IIa compounds, even though saturation solubility is approached in vivo, the formulation can
usually be designed to achieve complete oral absorption from a standard solid oral dosage form containing
crystalline drug without resorting to complex solubilisation technologies. This is a result of the compensatory influence of high permeability. However,
control of factors affecting drug release from a
standard formulation such as particle size, surface
area and wettability will be critical in achieving
complete absorption. In contrast, Class IIb compounds will tend to be incompletely absorbed unless
the drug is formulated in an already solubilised form,
and will therefore present a major challenge to the
formulator.
Note that, unlike the implication from the MAD
equation, the idea that permeability and solubility
are compensatory is not extended into class III. This
is because the revised classification system still
assumes a fixed initial volume available for dissolution in the intestine (500 mL) for both high and low
permeability compounds. In practice, the dynamic
nature of fluid secretion in, and absorption out of the
intestine may mean this is an over-simplification, but
an adequate one for the assessment of developability.
Also note that, due to the frequent use of multiple
dosage units during clinical studies during early
product development, it is reasonable to use the
maximum dose rather than the maximum dose
strength (as used for BCS classification) in the
modified classification system.
For convenience, the modified classification system
is referred to as the Developability Classification
System (DCS), to emphasise that it is aimed at
addressing issues in product development, rather
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 12, DECEMBER 2010

4944

BUTLER AND DRESSMAN

than being a regulatory classification system. However, biopharmaceutics still forms the rational basis
for the modified system.
For the incorporation of the concept that dissolution rate rather than dose/solubility ratio is more
important for dissolution rate limited drugs, the use
of a target dissolution number is advocated as
illustrated in Table 1. It demonstrates that via
rearrangement of Eqs. (4) and (5), it is possible to
use to calculate a target particle size below which the
particle size distribution of the drug needs to fall to
avoid dissolution rate (even under sink conditions)
becoming limiting to the extent of oral absorption.
This estimate will have most relevance for compounds
in class I, IIa, and III, as in class IIb and IV poor
solubility in the GI tract will limit oral absorption,
such that dose/solubility becomes the dominating
limiting factor. To illustrate the doses at which
these factors are predicted to become significant,
the solubility limited dose assuming a typical human
jejunal permeability of 2.0  104 cm/s (Absorption
number An 2) is also shown in Table 1.
Dissolution NumberDn 3D=r2 Cs=rTsi

(4)

where D is the diffusion coefficient (5  106 cm2/s), r


the particle radius (mm) and r is the drug density
(1.2 g/cm3).
r2 3D=DnCs=rTsi

(5)

Note that the small intestine transit time is


normally set to 3.32 h, but can be reduced if an
absorption window in the upper smaller intestine
is likely. Compounds that rely on active transport
mechanisms or on the paracellular route are those for
which a shorter absorption site transit time may be
useful.
It is important to note that the calculation is
primarily of use in determining the size control
necessary to avoid changes in the extent of absorption
of an immediate release product (reflected in vivo by
AUC). To ensure consistent Cmax, (indicative of rate of
absorption) particularly for low solubility/high permeability drugs, tighter control of particle size may be
Table 1. Particle Size Sensitivity of the Extent of Oral
Absorption of Dissolution Rate Limited Drugs Using a
Target Dissolution Number for a Target Dn 1
Drug Intestinal
Solubility (mg/mL)
0.1
0.05
0.01
0.005
0.001

Maximum
D90 Particle
Diameter (mm)

Solubility Limited
Absorbable Dose
Where An 2 (mg)

77.5
55
25
17.5
7.75

100
50
10
5
1

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 12, DECEMBER 2010

necessary, with the desired product performance (e.g.


whether rapid onset with high Cmax or prolonged
absorption with lower Cmax is preferred) also playing
a role in determining the desired product attributes
beyond complete oral absorption. The comparison
of dissolution profiles, in a manner similar to that
proposed by the BCS guidance, is likely to be
appropriate where there is a concern that formulation
change may alter the rate but not extent of
absorption.
Also note that when using a fixed single solubility
input based on the jejunal solubility, the accuracy of
the target particle size estimate will depend upon how
drug solubility is affected by pH changes in the upper
GI tract.
Accounting for Weak Bases
Adopting a single estimate of intestinal solubility as
the key solubility works well for poorly soluble
neutral drugs and weak acids, as the small intestine
is likely to be the primary site of in vivo dissolution.
However, due to their improved solubility at low
pH poorly soluble weak bases (typically with a pKa
between 3 and 9) have several additional factors
influencing oral absorption that need consideration:
(1) A high dependency upon gastric pH and residence time.
(2) A risk that drug will precipitate in the intestine
due to supersaturation.
Gastric pH and residence times are more variable
than in the intestine, and as a result, factors affecting
subject to subject variability such as feeding state,
disease state, age and use of other medications may
impact the oral absorption of poorly soluble weak
bases.
These factors mean that the use of a single
intestinal pH as the reference solubility equates to
worst case for weak base drug solubility, dissolution
and stomach residence time; that is where gastric pH
is high and transit through the stomach is rapid.
Typically though, there is a considerable upside on
this worst case due to some residence in a lower pH
environment favouring drug dissolution.
In practice, some weak bases may be highly reliant
on gastric dissolution (particularly where the dose
can fully dissolve under typical gastric conditions
without significant subsequent precipitation) whereas
others will be partly dependant on gastric and partly
on intestinal dissolution. For instance, Ketoconazole
is a drug known to have highly gastric pH dependent
oral bioavailability.48 Dipyridamole is a drug where
both gastric pH and precipitation in the upper
intestine may limit oral absorption from standard
IR dosage forms.49,50
DOI 10.1002/jps

THE DEVELOPABILITY CLASSIFICATION SYSTEM

In vitro test methods have been developed for the


prediction of precipitation of weak bases,50,51 but it
remains a problematic issue, as precipitation is a
more challenging property to predict in vitro than
dissolution rate. One simple approach is to consider
the likelihood of precipitation based on the difference
in gastric and intestinal solubility. For instance, if we
assume (for simplicity) that an equal fluid volume of
intestinal secretion is needed to neutralise fluid from
the stomach at a typical gastric pH, then it is at least
possible to assign to a given drug as to whether
precipitation is likely or not. For those drugs where
precipitation is likely, intestinal solubility is still a
highly significant factor for the redissolution of the
precipitated drug, albeit one with a great deal of
uncertainty due to the difficulty of predicting both
whether a drug precipitates in vivo and the nature of
any precipitated material. For drugs that do not
precipitate, gastric solubility alone may be adequate
to get the drug into solution in the GI tract for oral
absorption.
Factors that are likely to contribute to whether
precipitation occurs in the intestine include:
(a) The extent of supersaturation in the small
intestine, which is influenced by the relative
gastric to intestinal solubility and the rate of
gastric emptying.
(b) Physicochemical properties of the drug influencing supersaturation stability.
(c) The permeability of the drug.
(d) The presence/absence of precipitation inhibitors.
In terms of the modified classification system, the
application to poorly soluble weak bases therefore is
still relevant, but represents worst case. Weak bases
classified as IIa may not be particularly particle size
sensitive, and class IIb compounds may not always be
solubility limited, except in individuals with raised
gastric pH. More sophisticated modelling of the effect
of gastric pH and the subsequent in vivo pH gradient

4945

Figure 2. DCS categorisation of selected drugs.

on dissolution and precipitation may therefore be


necessary if an attempt at more rigorous assessment
of risks and potential CQAs is to be rigorously
performed.

RESULTS-SELECTED EXAMPLE DRUGS


The DCS categorisation of selected drugs is shown in
Figure 2. The classification of the selected drugs is
shown in Table 2.
Example 1: Digoxin 0.5 mg
Digoxin is a low dose, poorly soluble drug with a low
therapeutic index, and an interesting example of
where the current BCS could be regarded as
misleading. It has been used as a model compound
for dissolution rate limited absorption in the literature.15,52,53 The FaSSIF solubility of digoxin is
reported to be 0.017 mg/mL.54 An estimate of fraction
absorbed from solution of 81%55 was used to identify
digoxin as borderline for high/low permeability.
In spite of a favourable dose/solubility ratio, issues
with particle size dependent bioavailability have had
serious consequences for bioequivalence of this drug
in the past56,57 with micron-sized drug particles

Table 2. Properties of Selected Example Drugs


Drug and Dose
Digoxin 0.5 mg
Paracetamol 500 mg
Griseofulvin 500 mg
Mefenamic acid 250 mg
Ibuprofen 400 mg
Dipyridamole 100 mg
Acyclovir 800 mg
Furosemide 80 mg

BCS
Category

FaSSIF Solubility
(mg/mL)

Estimated Human Peff


(cm/s 104)

DCS
Category

SLADa
(mg)

Max Particle
Size (mm D90)

I/III
I
II
II
II
II
IV
IV

0.017
14
0.019
0.036
1.46
0.023
1.3
1.7

0.9
1.3
8.7
14
12
1.5
0.25
0.6

I/III
I
IIb
IIa
I
IIb
IV
III

8.5
9100
83
280
8800
17b
650
850

32
900
34
42
300
37b
280
120c

Solubility limited absorbable dose.


For dipyridamole, which is gastro-soluble, these intestinal solubility derived parameters will be of less significance.
For furosemide, the target particle size was calculated using a shorter intestinal transit time (0.5 h), assuming absorption is limited to the upper small
intestine.
b
c

DOI 10.1002/jps

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 12, DECEMBER 2010

4946

BUTLER AND DRESSMAN

needed for optimal drug absorption. Digoxin therefore


exemplifies a common misconception of the BCS
classification; that BCS I/III classification implies a
lack of dissolution rate sensitivity.
This does not necessarily mean the BCS I biowaiver
status of low dose but highly dissolution rate limited
drugs is inappropriate, as the high dependence on
dissolution rate at doses considerably below that for
saturation in the GI tract should enable statistically
similar dissolution in three media as recommended
in BCS guidance to ensure bio-equivalence. Even so,
highly soluble is a misleading term for such a drug.
One additional complication for digoxin is the
potential of this drug for oral absorption to be
modified by the inhibition of efflux transporters, in
particular p-glycoprotein (pgp). Evidence of this being
a factor influencing oral absorption of digoxin has
been shown using in situ perfusion in human
jejunum.58 The low dose (possibly combined with
low solubility and moderate permeability) makes
digoxin vulnerable to such effects. This may mean
such drugs carry some additional bio-equivalence
risk, although the relative significance of efflux
transporters in the gut to oral absorption is still a
matter of debate.40,59 For digoxin at least, the in vivo
evidence points to only small changes in the extent of
oral absorption in the presence of pgp inhibitors in
humans.60,61 It is possible that digoxin absorption is
limited to a greater extent in the lower small intestine
as a result of lower drug concentrations reaching the
distal small intestine and pgp efflux.62 If so, the 3.32 h
estimate of residence time at the absorption site used
to calculate the recommended maximum particle size
may be too generous, in which case even greater
sensitivity to particle size than indicated may be
anticipated.
In the revised classification system (see Fig. 2),
digoxin appears close to paracetamol in the plot
of dose/solubility against permeability. The critical
difference between these two drugs is solubilitydriven dissolution rate sensitivity, which is reflected
in the recommended maximum D90 particle sizes of
the two drugs: 32 mm for digoxin, but approaching
1000 mm for paracetamol (see Tab. 2).
Example 2: Griseofulvin 500 mg
Griseofulvin is a classic example of a solubilitylimited drug, used in previous papers modelling oral
absorption, often in comparison to the dissolution rate
limited drug digoxin.15,52,53 Solubility in FaSSIF
has been reported to be 19 mg/mL.63 Even though
solubility is similar to digoxin, there is not enough
fluid in the GI tract for complete dissolution, the dose
being three orders of magnitude greater than digoxin.
The permeability of griseofulvin is reported to be
high from cell lines.63 In addition absorption from
griseofulvin in solution in the human small intestine
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 12, DECEMBER 2010

has been shown to be rapid and unaffected by position


in the small intestine.64 For determining a permeability value for the modified classification, equivalent human jejunal permeability to ketoprofen was
assumed, based on the statistically similar published
Caco-2 permeability data for the two drugs.63
With the new DCS classification, even with an
estimate of very high permeability taken into
account, griseofulvin is clearly a solubility limited
drug. By quantifying the permeability in the DCS
model, it is possible to visualise how the good
permeability helps to partially mitigate low solubility,
enabling micron sized drug particles (below the
recommended maximum particle size) to be commonly used in oral formulations for griseofulvin to get
an adequate therapeutic response, even though oral
absorption is still incomplete and variable.
Confirmation that absorption is incomplete is
demonstrated by a solubility-related food effect for
doses of 1251000 mg65,66 even when the drug is
administered as micron sized particles. Administration of drug as GRIS PEG (a solid dispersion of small
drug particles in a polyethylene glycol) has been used
to improve the drug solubility to obtain greater oral
absorption compared to micronised drug in tablet
formulations.67 Getting complete oral absorption of
solubility limited drugs like griseofulvin is a significant challenge to the formulator, and the selection
of an appropriate means of solubilising the drug is
likely to be critical for optimal therapeutic performance. Such drugs may require additional resources
and time for formulation optimisation compared to
drugs that are not solubility-limited. A variety of
other formulation techniques for solubilisation of
griseofulvin, leading to improvement in oral absorption in humans are also reported in the literature.68,69
Example 3: Mefenamic Acid 250 mg
Mefenamic acid was one of the compounds used in
studies by Galia et al.19 in the establishment of
suitable biorelevant media for fasted and fed states.
Mefenamic acid is poorly soluble in both the proposed
fasted and fed state simulated intestinal media, with
only 57% of a 250 mg dose dissolving in 500 mL of
media during dissolution testing. The similar extent
of dissolution in vitro was linked to in vivo data on the
drug showing similar absorption fasted and fed.70
However, a weakness of this argument is that the
solubility of a weak acid with a pKa of 4.2 will be
highly pH dependent over the pH range typical of
the small intestine (pH 5.07.5), and pH 5.0, the pH
for FeSSIF is rather low as a pH for all but the
duodenal pH (at the very top of the small intestine)
experienced by the drug in the fed state intestine.
Mefenamic acid is a nonsteroidal anti-inflammatory drug (NSAID). Although permeability has not
been measured directly in a human jejunal perfusion
DOI 10.1002/jps

THE DEVELOPABILITY CLASSIFICATION SYSTEM

Table 3. Comparison of Properties of Three NSAIDS


Mefenamic Acid Ketoprofen Naproxen
log D (at pH 7.4)
pKa
Human Peff (cm/s 104)
a

2.069
4.2
14a

0.139
3.89
8.738

0.339
4.01
8.538

Estimated from structure and log D.

study, the permeability of several other structurally


similar NSAIDs, naproxen and ketoprofen, are
reported to have very high human jejunal permeability (see Tab. 3). Mefenamic acid is likely to be even
more permeable, as it is more lipophilic and has a
smaller fraction ionised in the intestine than ketoprofen or naproxen. Using an unpublished model
similar in concept that developed by Winiwarter
et al.,41 an estimate the passive permeability of
mefenamic acid was made. The model used log D and
several molecular descriptors known to be important
for permeability including polar surface area and
hydrogen bonding descriptors to predict human
jejunal permeability as being about 14  104 cm/s.
The structural similarity to compounds in the
measured data set increases the confidence in the
accuracy of this prediction.71
For estimating jejunal solubility in the revised
classification system, a value of 0.04 mg/mL was used,
derived from averaging two literature references
where mefenamic acid solubility in FaSSIF was
measured.19,72
Taking the estimate of jejunal permeability and of
FaSSIF solubility at pH 6.5 into the modified
classification system, it is then possible to derive an
alternative explanation for the observation of similar
oral absorption in humans fed and fasted. According
to the modified classification system, mefenamic acid
250 mg is a borderline IIa/IIb compound with a SLAD
of 280 mg and a recommended maximum D90 particle
size of 42 mm. Therefore, if dissolution rate is
sufficiently rapid, even from a formulation where
the drug is not present in a solubilised form, complete
absorption in the small intestine may be possible in
both fed and fasted states as high permeability
compensates for low solubility.
This explanation also helps to explain another
study comparing different mefenamic acid formulations in humans,73 where the food effect is shown to be
formulation dependent. Although the differences
between the formulations was not stated, it is possible
that measures to maximise the drug dissolution
rate from a standard tablet formulation such as
careful control of drug particle size and/or addition of
surfactant may be adequate for complete absorption
of mefenamic acid from the human GI tract.
Mefenamic acid and other DCS class IIa compounds
(carbamazepine and nitredipine are also shown as
DOI 10.1002/jps

4947

class IIa examples in Fig. 2) should be good


candidates for the development of in vitro/in vivo
relationships using dissolution rate as a predictor for
in vivo performance. Indeed, the possibility of such an
approach has been investigated in the literature.74,75
One additional interesting aspect of the study by
Shinkuma et al.74 provides further evidence that
dissolution rate is the primary barrier to oral
bioavailability. Milling of the capsule contents of
the formulation with the lowest AUC lead to
equivalent AUC to that obtained by the best
performing formulation in the study. However, the
high dependency of both AUC and Cmax on dissolution
rate, the high dose/solubility ratio and the proximity
of mefenamic acid to the class IIa/IIb boundary means
that the justification of biowaivers based on the
results of dissolution testing alone are probably
inappropriate.75
Example 4: Ibuprofen 400 mg
Ibuprofen is a BCS class II weak acid, with high
intestinal permeability, poor solubility in low pH
gastric environment but good solubility anticipated in
the small intestine.
FaSSIF solubility (1.46 mg/mL) taken from literature8 and human jejunal permeability was estimated
from the average of two independent literature rat
perfusion data sources,45,76 both of which indicated
ibuprofen has very high jejunal permeability, using
the relationships between human and rat jejunal
permeability already in the literature.45,46 The
average human jejunal permeability ( Peff) used in
the classification was 12  104 cm/s.
The new classification system assigns ibuprofen as
a DCS class I compound with extent of oral absorption
not affected by solubility. As such, it represents a
compound that should be relatively easily formulated
for optimal oral absorption. In addition, according to
the recommended maximum particle size, extent of
absorption is predicted to be only particle size
sensitive if drug particles exceed a D90 of 300 mm.
In vivo, literature evidence77 demonstrates that
ibuprofen is well absorbed, with significant formulation changes having little effect on the extent of oral
absorption (AUC).
As the DCS system is primarily concerned with
extent rather than rate of oral absorption and as
ibuprofen has poor gastric solubility, some sensitivity
in terms of rate and onset may be anticipated. In the
same study,77 the rate of absorption and onset of
action (illustrated by Cmax and Tmax) of ibuprofen in
the fasted state was altered via formulation.
Although Tmax and Cmax changed, significant formulation effort was still required for a 3555%
increase in Cmax. Note that the very high permeability
of ibuprofen may increase the vulnerability of the
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 12, DECEMBER 2010

4948

BUTLER AND DRESSMAN

drug to variation in Cmax. Less permeable drugs


should be less sensitive.
Although the main aim of the DCS is to assess
compounds in regard to their difficulty in formulating
to maximum oral absorption, DCS class I compounds,
even if not BCS I, may be suitable candidates for biowaivers. The case for ibuprofen has already been
highlighted in the literature.78 Even though there is
potential to change Cmax the risk of failing bioequivalence may be low provided similarity in dissolution
is demonstrated. Suitable in vitro dissolution profiles
combined with statistical similarity testing as outlined in the BCS may therefore be adequate to ensure
in vivo equivalence for BCS class II drugs classified as
DCS I.
In a review of ibuprofen clinical bio-equivalence
using regulatory data from Germany, it was found
that although bio-equivalence failures were relatively
common, theses were always related to Cmax rather
than AUC.79 It is unlikely that the products failing
equivalence underwent the rigorous dissolution
comparisons demanded by the BCS. This review also
found that there was a much weaker argument for
bio-waivers for mefenamic acid, due to vulnerability
to changes in both AUC and Cmax.
By choosing a highly permeable weak acid as a DCS
class I example, the worst case scenario is examined,
particularly in respect to Cmax. DCS I weak bases and
nonionised compounds will have higher relative
gastric solubility and therefore carry a lower risk of
bioequivalence failure. The DCS may therefore help
to define wider acceptance criteria for the extension of
BCS biowaivers into class II.
Example 5: Dipyridamole 100 mg
The solubility of dipyridamole (22.5 mg/mL) was taken
from a measurement in human aspirates.29 The
permeability of dipyridamole was estimated from the
same log D/molecular descriptor model used for
mefenamic acid. Although it was not possible to
make a quantitative assessment from measured
values, this estimate appears to be in line with data
generated in PAMPA80 and in Caco-2 cells.43
Dipyridamole is a drug with adequate gastric
solubility for dissolving the dose, but poor intestinal
solubility. As such, it represents a special case where
intestinal precipitation and gastric pH are likely to be
important for oral absorption. Despite the good
gastric solubility, the class IIb categorisation within
the DCS is appropriate, as formulating such drugs to
complete oral absorption in all patients may be very
challenging.
In vivo, dipyridamole does show some variability in
bioavailability. Gastric pH related variability has
been demonstrated in elderly patients81 and raised
gastric pH significantly lowers the oral absorption of
the drug.82 Dosing as a cyclodextrin formulation
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 12, DECEMBER 2010

helps to reduce variability and increased bioavailability.83 In addition, the use of acidifying agents in
an extended release formulation has been used to
improve the pharmacokinetic variability.49,82 This
approach probably works by controlling the microenvironmental pH around the dissolving drug.
Dipyidamole exemplifies the issues sometimes
faced with poorly soluble weak bases. Weak bases
that are solubility limited in the small intestine (class
IIb or IV) present a particular challenge, as bioavailability is gastric pH dependent, and intestinal
precipitation is a potential risk. This risk is expected
to be much reduced for weak bases in class IIa due to
the compensatory influence of high permeability and
negligible those in class I/III.
Example 6: Furosemide 80 mg
Furosemide is a BCS class IV diuretic. It shows
variable oral absorption with several reports indicating oral absorption being limited mainly to the upper
GI tract.84,85 Despite BCS class IV status, it is
generally available as a standard immediate release
tablet indicating that relatively simple formulation
techniques are generally adequate for reasonable oral
absorption.86 Some improvement in the oral absorption of furosemide can be obtained via mucoadhesion
to keep the drug in the upper GI tract for longer.87
The absorption window for furosemide may in part be
due to higher luminal pH in the lower small intestine,
with more extensive ionisation leading to reduced
permeability. The use of an acidified suspension
shows improved absorption, at least in rats.85 Active
efflux of furosemide in the GI tract according to one
report,88 may additionally contribute to variability
and the absorption window for oral absorption in the
upper small intestine.
As furosemide is a weak acid, it is the low gastric
solubility that drives the low solubility classification
according to BCS. In contrast, good intestinal
solubility (1.7 mg/mL in FaSSIFsee Tab. 4) and
moderate dose (2080 mg) means furosemide can be
placed into DCS class III.
Development of an appropriate formulation for
furosemide does not require specialist formulation
techniques as the drug can be adequately delivered
from a simple oral formulation. However, the limited
Table 4. Solubility of Low Permeability Drugs in
FaSSIF pH 6.5

Acyclovir
Ganciclovir
Chlorothiazide
Furosemide

Saturation Solubility (mg/mL)

Final pH

1.3
2.6
0.3
1.7

6.3
6.4
6.4
5.9

Determined at room temperature.


Dissolved furosemide significantly altered media pH.
DOI 10.1002/jps

THE DEVELOPABILITY CLASSIFICATION SYSTEM

absorption in the intestine and low gastric solubility


means that there is some variability in the oral
absorption depending on dosing and formulation. It
would appear from the literature that these sensitivities mainly relate to formulation parameters that
affect dissolution rate,86,8991 with dissolution media
of pH 46 being preferred for optimal discrimination
of in vivo performance.86,89,90 However, changes to
fluid volume or upper intestinal transit time also
appear influential.92 In addition, food tends to lead to
some reduction in AUC,93 which is typical for
permeability rather than solubility limited drugs.
Therefore, the BCS IV categorisation for biowaivers
and DCS III for developability may both be appropriate for furosemide.
A modification in the calculation of the recommended maximum particles size may be necessary
to adequately reflect the shorter time for optimal
absorption in the upper GI tract. In Table 2, a reduced
absorption site transit time of 30 min is used in Eq. (5)
rather than the usual 3.3 h to calculate the maximum
particle size to reflect a narrow absorption window.
There is some evidence that for a significant
proportion of poorly permeable drugs, permeability
decreases with progression along the GI tract, at least
in rats,94 so assuming a shorter effective transit time
when predicting oral absorption may be a sensible
precaution for poorly permeable compounds.
The Class III/IV Boundary
The validity of a single 500 mL fluid volume in the
model can be rationalised by examining pharmacokinetic data on drugs at the class III/IV boundary.
Poorly permeable drugs nonionised across the
physiological pH range such as acyclovir and gancyclovir have human dose/solubility ranges that cross
the BCS/DCS 250/500 mL boundary (see Fig. 2),
whilst the weak acid chlorothiazide human dose
range is in DCS class IV. By examining the known
biopharmaceutical performance of these drugs, it is
possible to identify at what point solubility (rather
than permeability) related issues become a serious
developability issue.
All three drugs show dose dependent bioavailability, with higher extent of oral absorption at lower
doses. Gancyclovir appears to show dose dependent
absorption, with a significant drop in dose normalised
absorption at 2000 mg compared with 1000 mg.95
Acyclovir bioavailability drops from about 2030% at
200 mg, to about 1012% at 800 mg, with 400 mg
being about 1521%.9698 Chlorothiazide shows a
strongly dose dependant oral absorption across a wide
range of doses, attributed to saturation of an uptake
process and/or solubility limited absorption.99 250 mg
and 500 mg doses seem to give similar exposure levels
despite the doubling of dose.100 Changes in bioavailability with dose for all three of these compounds may
DOI 10.1002/jps

4949

be due to saturation of active uptake rather than


solubility, as highlighted as a risk for class III
compounds by Wu and Benet.10 This makes it difficult
to identify a critical volume for solubility limited
absorption from dose escalation data alone, although
the apparent plateau in exposure with dose seen with
chlorothiazide may be taken as an indicator of
solubility limited absorption at 250 mg and above.
In regard to food effects, gancyclovir is known to
exhibit a small (20% increase) but significant food
effect at 1000 mg,101 but in a separate study, showed a
larger, more significant doubling of AUC with food
at 2000 mg.102 At conventional doses acyclovir
(3  200 mg dosed 4 h apart) shows equivalent AUCs
fasted and fed,98 although a high fat meal has been
shown to slightly reduce (by about 18%) the extent of
oral absorption of a 400 mg dose compared to a low fat
meal.103 Interestingly, in an unpublished study at a
higher dose (800 mg) acyclovir was more (almost
twofold) extensively absorbed after food, and also
appears to be more extensively absorbed if dosed with
500 mL rather than a 50 mL volume of water.104
Chlorothiazide shows a doubling of AUC with food at
500 mg.105 Assuming small changes in AUC may be
related to either saturation of active uptake or
solubility, but large increases in AUC with food are
highly likely to be related to inadequate fluid volume
in the GI tract, then chlorothiazide 500 mg, acyclovir
800 mg, and ganciclovir 2000 mg appear to be
solubility limited, in contrast to acyclovir 400 and
1000 mg ganciclovir, which could be described as
borderline.
The solubility of these three drugs, along with
furosemide was measured in FaSSIF. The results are
shown in Table 4. These solubility values together
with selected doses and permeability estimated from
fraction absorbed at low dose enabled the drugs to
be represented in Figure 2. A best fit for these three
drugs gives the boundary for solubility limited
absorption in the order of 300500 mL.

DISCUSSION
The BCS is used in the pharmaceutical industry for
developability assessment, and is attractive due to its
conceptual simplicity. However, based on the evidence presented, the DCS is a convenient way of
classifying compounds in a way that is more meaningful in determining human in vivo behaviour and
sensitivities. In addition, although other more complex models such as the commercial software GastroPlus and PK-Sim may also useful in assessing drug
developability issues, their optimal use requires some
expertise in biopharmaceutics and mathematical
modelling, and lacks the accessibility and simplicity
of a classification system. Therefore the revised
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 12, DECEMBER 2010

4950

BUTLER AND DRESSMAN

classification system fills a gap for a simple, but more


realistic simple classification for developability analysis.
The ideas presented here do not represent so much
a refinement of current models, but a way of
assembling the most useful concepts into a simple
classification system. As such, the DCS may be
particularly useful in explaining the risks and
challenges in the development of oral products
to a wider audience than those individuals with
a specialist knowledge of biopharmaceutics and
modelling.
One key advantage for the DCS is that it gives the
formulator an early warning as to which poorly
soluble drugs can be adequately formulated via
simple size control versus which are likely to need
specialist solubilisation techniques to obtain complete
oral absorption and to avoid solubility related food
effects.
Although not discussed in detail here, a drugs
location within the solubility limited (IIb/IV) region
may be an indicator of the formulation approach most
likely to successfully solubilise the drug. Drugs
optimally delivered either by nanoparticles, solid
dispersions or lipidic formulations are likely to occupy
different regions within the solubility limited (IIb/IV)
space. Ultimately, if the position in this region is
highly unfavourable, the challenge of obtaining
adequate solubilisation in the GI tract may be such
that redesign of the molecule becomes the more
prudent strategy. Knowledge of the most likely
barriers to good oral absorption early in development
means it is then possible to introduce appropriate
formulation strategies for early clinical studies.
Although preclinical use of animal species can also
be used for the identification of potentially solubility
limited drugs, physiological and metabolic differences
between different species can make translation to
human performance problematic. In practice, the use
of animal studies and modelling/classification tools
are best used as complementary methods in developability assessment.
Limitations and Assumptions
In devising the modified classification system, several
assumptions were necessary for sake of simplicity.
These include:
(1) Single values for upper small intestine solubility and permeability assumed to adequately
represent solubility and permeability in the
region of interest for oral absorption.
(2) A fluid volume of about 500 mL is a reasonable
approximation for the total volume of fluid
available in the GI tract in the fasted state
for drug dissolution.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 12, DECEMBER 2010

(3) Permeability and solubility are compensatory


for highly permeable/low solubility compounds.
(4) Dissolution rate and solubility are assumed to
be directly related in the simple manner
expressed in the dissolution number (Dn) equation (Eqs. 4 and 5).
The revised classification system will be most
reliable when a drug is nonionised in the physiological pH range. For ionised drugs, the pH solubility
profile and the permeability profile down the GI tract
cannot be fully represented by single values. However, by choosing to estimate solubility at a point in
the GI tract of most significance for oral absorption,
the effect of this simplification is minimised. This is
most likely to be an over-simplification for poorly
soluble weak bases, as previously outlined. This is a
disadvantage of the DCS compared to the use of more
complex modelling tools such as Gastro-Plus.
The estimation of intestinal solubility via the use of
biorelevant fluids means that the true solubility of
lipophilic drugs in the GI tract can be more accurately
estimated. However, there will be some margin of
error between this estimate and the actual in vivo
solubility. Differences can be anticipated to be due to
the ability of simulated media to use the exact buffer
components and buffer capacity106 compared to that
present in the intestine. In addition, bile salts and
other solubilising agents in the intestine are a more
complex mixture than that present in simulated
fluids. These simplifications mean that solubility of
high doses of some drugs in vivo may deviate from
in vitro measurement of solubility due to their
influence on local pH,106 and highly lipophilic drugs
may not be truly represented by simulated media
solubility.24
Even if intestinal solubility can be accurately
estimated, dissolution rate may not follow the
simplistic relationship to solubility implied by the
dissolution number equation, particularly when
the drug approaches saturation solubility or is
significantly solubilised in bile salt micelles. This is
most likely to be of significance for highly lipophilic
drugs and in the fed state, where micellar solubilisation may be more extensive.107 The estimate of a
maximum recommended particle size may therefore
lead to some over-estimation for highly lipophilic
drugs, as will the assumption of sink conditions in the
determination of the recommended particle size. The
sink condition assumption should hold for drugs that
are DCS class I or III, but may be an underestimate
for size effects on solubility limited drugs (IIb and IV)
and to a lesser extent for class IIa drugs that are likely
to approach saturation concentrations at least at
some point in the GI tract.
Another assumption of the dissolution number
equation, the 3.3 h transit time of the small intestine
DOI 10.1002/jps

THE DEVELOPABILITY CLASSIFICATION SYSTEM

as the maximum time over which dissolution can take


place, may also be too generous for compounds that
are regionally absorbed in only part of the small
intestine. Drugs that rely on active transport
processes for their absorption and those that are
poorly absorbed are at a higher risk of such an
absorption window. For these drugs, it may be better
to assume a shorter transit time in any model of oral
absorption, as illustrated for furosemide.
Similarly, significant deviations between actual
and estimated permeability in the GI tract will also
introduce some degree of error into the classification
system. The difficulties in estimating the true
intestinal solubility and permeability means that
further evidence (e.g. from dosing orally to preclinical
species) is desirable when making decisions about the
most appropriate formulation for early clinical work.

CONCLUSION
Based on the examples shown, a revised version of
the BCS for assessing developability of drugs for
oral immediate release delivery is a useful way
of categorising compounds in a simple manner to
identify whether dose/solubility ratio, dissolution rate
and/or permeability are likely to limit oral absorption
of a drug. The introduction of a target particle size via
rearrangement of the dissolution number equation
provides additional useful information on potential
dissolution rate sensitivity. In addition, the concept
of solubility limited absorbable dose (SLAD) may
improve the estimate of the dose above which oral
absorption is likely to be limited by intestinal
solubility.
The visualisation of which factors are most likely
to put the in vivo performance of a drug at risk is
also helpful in the context of defining strategies
for optimal formulation and highlighting potential
critical quality attributes (CQAs).

REFERENCES
1. Amidon GL, Lennernas H, Shah VP, Crison JR. 1995. A
theoretical basis for a biopharmaceutic drug classification
system: The correlation of in vitro drug product dissolution
and in vivo bioavailability. Pharm Res 12:413420.
2. FDA Guidance: Waiver of in vivo bioavailability and bioequivalence studies for immediate-release solid oral dosage forms
based on a biopharmaceutics classification system. Issued
Aug. 2000, see http://www.fda.gov/cder/guidance/.
3. ICH guidance documents Q8 R1 2009 (Pharmaceutical Development Revision 1), Q9 2006 (Quality Risk Management) and
Q10 2009 (Pharmaceutical Quality System) at http://
www.fda.gov/cder/guidance/.
4. Basit A, Podczeck F, Newton M, Waddington W, Ell P, Lacey
L. 2002. Influence of polyethylene glycol 400 on the gastrointestinal absorption of ranitidine. Pharm Res 19:13681374.
DOI 10.1002/jps

4951

5. Adkin DA, Davis SS, Sparrow RA, Huckle PD, Wilding IR.
1995. The effect of mannitol on the oral bioavailability of
cimetidine. J Pharm Sci 84:14051409.
6. Yu LX, Amidon GL, Polli J, Zhao H, Mehta M, Conner D, Shah
VP, Lesco LJ, Chen ML, Lee VHL, Hussain AS. 2002. Biopharmaceutics classification system: The scientific basis for
biowaiver extensions. Pharm Res 19:921925.
7. Cheng CL, Yu LX, Lee HL, Yang CY, Lue CS, Chou CH. 2004.
Biowaiver extension potential to BCS Class III high solubilitylow permeability drugs: Bridging evidence for Metformin
immediate-release tablet. Eur J Pharm Sci 22:297304.
8. Yazdanian M, Briggs K, Jankovsky C, Hawi A. 2004. The
high solubility definition of the current FDA guidance on
biopharmaceutical classification system may be too strict for
acidic drugs. Pharm Res 21:293299.
9. WHO 2006. Proposal to waive in vivo bioequivalence requirements for the WHO Model List of Essential Medicines immediate release, solid oral dosage forms. In: Fortieth report of the
WHO Expert Committee on Specifications for Pharmaceutical
Preparations. Geneva, World Health Organization. Technical
Report Series, no. 937, Annex 8.
10. Wu CY, Benet LZ. 2005. Predicting drug disposition via
application of BCS: Transport/absorption/elimination interplay and development of a Biopharmaceutics Drug Disposition
Classification System. Pharm Res 22:1123.
11. Curatolo W. 1998. Physical chemical properties of oral drug
candidates in the discovery and exploratory development
settings. Pharm Sci Techol Today 1:387393.
12. Sun D, Yu LX, Hussain MA, Wall DA, Smith RL, Amidon GL.
2004. In vitro testing of drug absorption for drug developability assessment. Cur Opin Drug Discov Dev 7:7585.
13. Fagerholm U, Nilsson D, Knutson L, Lennernas H. 1999.
Jejunal permeability in humans in vivo and rats in situ:
Investigation of molecular size selectivity and solvent drag.
Acta Physiol Scand 165:315324.
14. Martinez MN, Amidon GL. 2002. A mechanistic approach to
understanding the factors affecting drug absorption: A review
of fundamentals. J Clin Pharmacol 42:620643.
15. Yu LX, Lipka E, Crison JR, Amidon GL. 1996. Transport
approaches to the biopharmaceutical design of drug delivery
systems: Prediction of intestinal absorption. Adv Drug Deliv
Rev 19:359376.
16. Yu LX. 1999. An integrated model for determining causes of
poor oral drug absorption. Pharm Res 16:18831887.
17. Willmann S, Schmitt W, Keldenich J, Lippert J, Dressman J.
2004. A physiological model for the estimation of the fraction
dose absorbed in humans. J Med Chem 47:40224031.
18. Dressman JB, Amidon GL, Reppas C, Shah VP. 1998. Dissolution testing as a prognostic tool for oral drug absorption:
Immediate release dosage forms. Pharm Res 15:1122.
19. Galia E, Nicolaides E, Horter D, Lobenberg R, Reppas C,
Dressman JB. 1988. Evaluation of various dissolution media
for predicting in vivo performance of class I and II drugs.
Pharm Res 15:698705.
20. Vertzoni M, Dressman JB, Butler JM, Hempenstall J, Reppas
C. 2005. Simulation of fasting gastric conditions and its
importance for the in vivo dissolution of lipophilic compounds.
Eur J Pharm Biopharm 60:413417.
21. Jantratid E, Janssen N, Reppas C, Dressman JB. 2008. Dissolution media simulating conditions in the proximal human
gastrointestinal tract: An update. Pharm Res 25:16631676.
22. Porter C, Charman W. 2001. In vitro assessment of oral lipid
based formulations. Adv Drug Deliv Rev 50:S127S147.
23. Dressman JB, Butler JM, Hempenstall J, Reppas C. 2001. The
BCS: Where do we go from here? Pharm Tech 25:6876.
24. Vertzoni M, Fotaki N, Kostewicz E, Stippler E, Leuner C,
Nicolaides E, Dressman J, Reppas C. 2004. Dissolution media
simulating the intralumenal composition of the small intesJOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 12, DECEMBER 2010

4952

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.
41.

42.

BUTLER AND DRESSMAN

tine: Physiological issues and practical aspects. J Pharm


Pharmacol 56:453462.
Kalantzi L, Persson E, Polentarutti B, Abrahamsson B, Goumas K, Dressman JB, Reppas C. 2006. Canine intestinal
contents vs. simulated media for the assessment of solubility
of two weak bases in the human small intestinal contents.
Pharm Res 23:13731381.
Fadda HM, Basit AW. 2005. Dissolution of pH responsive
formulations in media resembling intestinal fluids: Bicarbonate versus phosphate buffers. J Drug Deliv Sci Techol 15:
273279.
Vertzoni M, Pastelli E, Psachoulias D, Kalantzi L, Reppas C.
2007. Estimation of intragastric solubility of drugs: In what
medium? Pharm Res 24:909917.
Persson EM, Gustafsson AS, Carlsson AS, Knutson L,
Hanisch G, Lennernas H, Abrahamsson B. 2005. The effects
of food on the dissolution of poorly soluble drugs in human and
in model small intestinal fluids. Pharm Res 22:21412151.
Kalantzi L, Goumas K, Kalioras V, Abrahamsson B, Dressman JB, Reppas C. 2006. Characterization of the human
upper gastrointestinal contents under conditions simulating
bioavailability/bioequivalence studies. Pharm Res 23:165
176.
Dressman JB, Vertzoni M, Goumas K, Reppas C. 2007. Estimating drug solubility in the gastrointestinal tract. Adv Drug
Deliv Rev 59:591602.
Bergstrom C, 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:
558570.
Abraham M, Ibrahim A, Zissimos AM, Zhao YH, Comer J,
Reynolds DP. 2002. Application of hydrogen bonding calculations in property based drug design. Drug Discov Today 7:
10561063.
Zhu C, Jiang L, Chen TM, Hwang KK. 2002. A comparative
study of artificial membrane permeability assay for high
throughput profiling of drug absorption potential. Eur J
Med Chem 37:399407.
Irvine J, Takahashi L, Lockhart K, Cheong J, Tolan J, Selick
E, Grove R. 1999. MDCK cells: A tool for membrane permeability screening. J Pharm Sci 88:2833.
Usansky H, Sinko P. 2005. Estimating human drug oral
absorption kinetics from Caco-2 permeability using an absorption-disposition model: Model development and evaluation
and derivation of analytical solutions for ka and Fa. J Pharmacol Exp Ther 314:391399.
Tavelin S, Taipalensuu J, Hallbook F, Vellonen KS, Moore V,
Artursson P. 2003. An improved cell culture model based on
2/4/a1 cell monolayers for studies of intestinal drug transport:
Characterization of transport routes. Pharm Res 20:373381.
Ruan LP, Chen S, Yu BY, Zhu DN, Cordell GA, Qiu SX. 2006.
Prediction of human absorption of natural compounds by the
non-everted rat intestinal sac model. Eur J Med Chem 41:605
610.
Levet-Trafit B, Gruyer MS, Marjanovic M, Chou RC. 1996.
Estimation of oral drug absorption in man based on intestine
permeability in rats. Life Sci 58:359363.
Lennernas H. 2007. Animal data: The contributions of the
using chamber and perfusion systems to predicting human
oral drug delivery in vivo. Adv Drug Deliv Rev 59:11031120.
Lennernas H. 2007. Intestinal permeability and its relevance
for absorption and elimination. Xenobiotica 37:10151051.
Winiwarter S, Bonham N, Ax F, Hallberg A, Lennernas H,
Karlen A. 1998. Correlation of human jejunal permeability of
drugs with experimentally and theoretically derived parameters. J Med Chem 41:49394949.
Winiwarter S, Ax F, Lennernas H, Hallberg A, Pettersson C,
Karlen A. 2003. Hydrogen bonding descriptors in the predic-

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 12, DECEMBER 2010

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.
58.

59.

60.

61.

tion of human in vivo intestinal permeability. J Mol Graph


Model 21:273287.
Sugano K, Nabuchi Y, Machida M, Aso Y. 2003. Prediction of
human intestinal permeability using artificial membrane
permeability. Int J Pharm 257:245251.
Sun D, Lennernas H, Welage L, Barnett J, Landowski C,
Foster D, Fleisher D, Lee KD, Amidon GL. 2002. Comparison
of human duodenum and Caco-2 gene expression profiles for
12,000 gene sequences tags and correlation with permeability
of 26 drugs. Pharm Res 19:14001415.
Zakeri-Milani P, Valizadeh H, Tajerzadeh H, Azarmi Y, Islambolchilar Z, Barzegar S, Barzegar-Jalali M. 2007. Predicting
human intestinal permeability using single-pass intestinal
perfusion in rat. J Pharm Pharm Sci 10:368379.
Fagerholm U, Johansson M, Lennernas H. 1996. Comparison
between coefficients in rat and human jejunum. Pharm Res
13:13361342.
Salphati L, Childers K, Pan L, Tsutsui K, Takahashi L. 2001.
Evaluation of a single perfusion method in rat for the prediction of absorption in man. J Pharm Pharmacol 53:10011013.
Van der Meer JWM, Keuning JJ, Scheijgrond HW, Heykants
J, Van Cutsem J, Brugmans J. 1980. The influence of gastric
acidity on the bioavailability of ketoconazole. J Antimicrobial
Chemother 6:552554.
Derendorf H, VanderMaelen C, Brickl R, MacGregor T, Eisert
W. 2005. Dipyridamole bioavailability in subjects with
reduced gastric acidity. J Clin Pharmacol 45:845850.
Kostewicz ES, Wunderlich M, Brauns U, Becker R, Bock T,
Dressman JB. 2004. Predicting the precipitation of poorly
soluble weak bases upon entry in the small intestine. J Pharm
Pharmacol 56:4351.
Gu CH, Rao D, Gandhi R, Hilden J, Raghavan K. 2004. Using
a novel multicompartment dissolution system to predict the
effect of gastric pH on the oral absorption of weak bases with
poor intrinsic solubility. J Pharm Sci 94:199208.
Dressman J, Fleisher D. 1986. Mixing-tank model for predicting dissolution rate control of oral absorption. J Pharm Sci
75:109116.
Rinaki E, Dokoumetzidis A, Macheras P. 2003. The mean
dissolution time depends on the dose/solubility ratio. Pharm
Res 20:406408.
Alsenz J, Meister E, Haenel E. 2007. Development of a partially automated solubility screening (pass) assay for early
drug development. J Pharm Sci 96:17481762.
Hinderling P, Hartmann D. 1991. Pharmacokinetics of
digoxin and main metabolites/derivatives in healthy humans.
Ther Drug Monit 13:381401.
Greenblatt DJ, Smith TW, Koch-Weser J. 1976. Bioavailability of drugs: The digoxin dilemma. Clin Pharmacokinet 1:
3651.
Jounela AJ, Sothmann A. 1973. Bioavailability of digoxin.
Lancet 1:202203.
Igel S, Drescher S, Murdter T, Hofmann U, Heinkele G,
Tegude H, Glaeser H, Bronner S, Somogyi A, Omori T, SchA
fer C, Eichelbaum M, Fromm M. 2007. Increased absorption of
digoxin from the human jejunum due to inhibition of intestinal
transporter-mediated efflux. Clin Pharmacokinet 46:777785.
Hunter J, Hirst B. 1997. Intestinal secretion of drugs. The role
of P-glycoprotein and related drug efflux systems in limiting
oral drug absorption. Adv Drug Deliv Rev 25:129157.
Westphal K, Weinbrenner A, Giessmann T, Stuhr M, Franke
G, Zschiesche M, Oertel R, Terhaag B, Kroemer HK, Siegmund W. 2000. Oral bioavailability of digoxin is enhanced by
talinolol: Evidence for involvement of intestinal P-glycoprotein. Clin Pharmacol Ther 68:612.
Verstuyft C, Strabach S, El-Morabet H, Kerb R, Brinkmann
U, Dubert L, Jailon P, Funck-Brentano C, Trugnan G,
Becquemont L. 2003. Dipyridamole enhances digoxin
DOI 10.1002/jps

THE DEVELOPABILITY CLASSIFICATION SYSTEM

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

bioavailability via P-glycoprotein inhibition. Clin Pharmacol


Ther 73:5160.
Sababi M, Borga O, Hultkvist-Bengtsson U. 2001. The role of
P-glycoprotein in limiting intestinal absorption of digoxin in
rats. Eur J Pharm Sci 14:2127.
Takano R, Sugano K, Higashida A, Hayashi Y, Machida M,
Aso Y, Yamashita S. 2006. Oral absorption of poorly watersoluble drugs: Computer simulation of fraction absorbed in
humans from a miniscale dissolution test. Pharm Res 23:
11441155.
Gramatte T. 1994. Griseofulvin. Absorption from different
sites in the human small intestine. Biopharm Drug Dispos
15:747759.
Ogunbona FA, Smith IF, Olawoye OS. 1985. Fat contents of
meals and bioavailability of griseofulvin in man. J Pharm
Pharmacol 37:283284.
Crounse RG. 1961. Human pharmacology of griseofulvin: The
effect of fat intake on gastrointestinal absorption. J Invest
Dermatol 37:529533.
Schafer-Korting M, Korting HC, Mutschler E. 1985. Human
plasma and skin blister fluid levels of griseofulvin following a
single oral dose. Eur J Clin Pharmacol 29:109113.
Ahmed IS, Aboul-Einien MH. 2007. In vitro and in vivo
evaluation of a fast-disintegrating lyophilized dry emulsion
tablet containing griseofulvin. Eur J Pharm Sci 32:5868.
Bates T, Sequeira J. 1975. Bioavailability of micronised griseofulvin from corn oil-in-water emulsion, aqueous suspension, and commercial tablet dosage forms in humans. J Pharm
Sci 64:793797.
Hamaguchi T, Shinkuma D, Yamanaka Y, Mizuno N. 1986.
Bioavailability of mefenamic acid: Influence of food and water
intake. J Pharm Sci 75:891893.
Mukherjee A, Hale VG, Borga O, Stein R. 1996. Predictability
of the clinical potency of NSAIDs from the preclinical pharmacodynamics in rats. Inflam Res 45:531540.
Devani M, Ashford M, Craig D. 2004. The emulsification
and solubilisation properties of polyglycolysed oils in selfemulsifying formulations. J Pharm Pharmacol 56:307
316.
Hamaguchi T, Shinkuma D, Yamanaka Y, Mizuno N. 1987.
Effects of food on absorption of mefenamic acid from two
commercial capsules differing in bioavailability under the
fasting state. J Pharmacobio-Dynam 10:2125.
Shinkuma D, Hamaguchi T, Yamanaka Y, Mizuno N. 1984.
Correlation between dissolution rate and bioavailability of
different commercial mefenamic acid capsules. Int J Pharm
21:187200.
Hummel D, Buchmann S. 2000. Einfluss der teilchengrosse
von mefenaminsaure auf dissolution und bioverfugbarkeit
von tabletten. Pharmazeutische Industrie 62:452456.
Levis K, Lane M, Corrigan O. 2003. Effect of buffer media
composition on the solubility and effective permeability coefficient of ibuprofen. Int J Pharm 253:4959.
Schettler T, Paris S, Pellett M, Kidner S, Wilkinson D. 2001.
Comparative pharmacokinetics of two fast-dissolving oral
ibuprofen formulations and a regular-release ibuprofen tablet
in healthy volunteers. Clin Drug Investig 21:7378.
Potthast H, Dressman JB, Junginger HE, Midha KK, Oeser H,
Shah VP, Vogelpoel H, Barends DM. 2005. Biowaiver monographs for immediate release solid oral dosage forms: Ibuprofen. J Pharm Sci 94:21212131.
Langguth P, Bolger M, Tubic M. Biowaiver for BCS class II
compounds? Application of simulation for BCS classification.
AAPS workshop on BE, BCS, and beyond. North Bethesda,
MD 2007 May 2123.
Bendels S, Tsinman O, Wagner B, Lipp D, Parrilla I, Kansy M,
Avdeef A. 2006. PAMPA-excipient classification gradient map.
Pharm Res 23:25252535.

DOI 10.1002/jps

4953

81. Russell TL, Berardi RR, Barnett JL, OSullivan TL. 1994. pH
related changes in the absorption of dipyridamole in the
elderly. Pharm Res 11:136143.
82. Kohri N, Miyata N, Takahashi M, Endo H, Iseki K, Miyazaki
K, Takechi S, Nomura A. 1992. Evaluation of pH-independent
sustained-release granules of dipyridamole by using gastricacidity-controlled rabbits and human subjects. Int J Pharm
81:4958.
83. Ricevuti G, Mazzone A, Pasotti D, Uccelli E, Pasquali F,
Gazzani G, Fregnan G. 1991. Pharmacokinetics of dipyridamole beta cyclodextrin in healthy volunteers after single and
multiple doses. Eur J Drug Metab Pharmacokinet 16:197
201.
84. Riley SA, Kim M, Sutcliffe F, Rowland M, Turnberg LA.
1992. Absorption of polar drugs following caecal instillation
in healthy volunteers. Aliment Pharmacol Ther 6:701
706.
85. Terao T, Matsuda K, Shouji H. 2001. Improvement in sitespecific intestinal absorption of furosemide by Eudragit L10055. J Pharm Pharmacol 53:433440.
86. Kingsford M, Eggers N, Soteros G, Maling T, Shirkey R. 1984.
An in-vivoin-vitro correlation for the bioavailability of frusemide tablets. J Pharm Pharmacol 36:536538.
87. Akiyama Y, Nagahara N, Nara E, Kitano M, Iwasa S, Yamamoto I, Azurma J, Ogawa Y. 1998. Evaluation of oral mucoadhesive microspheres in man on the basis of the
pharmacokinetics of furosemide and riboflavin, compounds
with limited gastrointestinal absorption sites. J Pharm Pharmacol 50:159166.
88. Flanagan SD, Benet LZ. 1999. Net secretion of furosemide is
subject to indomethacin inhibition, as observed in Caco-2
monolayers and excised rat jejunum. Pharm Res 16:221224.
89. Prasad V, Rapaka R, Knight P, Cabana B. 1982. Dissolution
medium: A critical parameter to identify bioavailability problems of furosemide tablets. Int J Pharm 11:8190.
90. McNamara PJ, Foster TS, Digenis GA, Patel RB, Craig WA,
Welling PG, Rapaka RS, Prasad VK, Shah VP. 1987. Influence
of tablet dissolution on furosemide bioavailability: A bioequivalence study. Pharm Res 4:150153.
91. Rubenstein MH, Rughani JM. 1978. Effect of four tablet
binders on the bioavailability of frusemide from 40 mg tablets.
Drug Dev Ind Pharm 4:541553.
92. Molz KH, Pabst G, Dilger C, Weber W, Renner P, Jaeger H.
1991. Multiple peaks and low bioavailability of furosemide
correlate with the volume of fluid ingested. Eur J Drug Met
Pharmacokinet Spec No. 3:194200.
93. Beermann B, Midskov C. 1986. Reduced bioavailability and
effect of furosemide given with food. Eur J Clin Pharmacol
29:725727.
94. Ungell A, Nylander S, Bergstrand S, SjoBerg A, Lennernas H.
1998. Membrane transport of drugs in different regions of the
intestinal tract of the rat. J Pharm Sci 87:360366.
95. Yang Z, Manitpisitkul P, Sawchuk R. 2006. In situ studies of
regional absorption of lobucavir and ganciclovir from rabbit
intestine and predictions of dose-limited absorption and associated variability in humans. J Pharm Sci 95:22762292.
96. Vergin H, Kikuta C, Mascher H, Metz R. 1995. Pharmacokinetics and bioavailability of different formulations of aciclovir.
Arzneimittel-Forschung/Drug Res 45:508515.
97. Laskin OL. 1983. Clinical pharmacokinetics of acyclovir. Clin
Pharmacokinet 8:187201.
98. Prescribing Information for Zovirax (acyclovir, GlaxoSmithKline). http://us.gsk.com/products/assets/us_zovirax.pdf.
99. Hsu F, Prueksaritanont T, Lee M, Chiou W. 1987. The phenomenon and cause of the dose-dependent oral absorption of
chlorothiazide in rats: Extrapolation to human data based on
the body surface area concept. J Pharmacokinet Biopharm
15:369386.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 12, DECEMBER 2010

4954

BUTLER AND DRESSMAN

100. Shah V, Knight P, Vadlamani KP, Cabana B, Thiazides IV.


1982. Comparison of dissolution with bioavailability of chlorothiazide tablets. J Pharm Sci 71:822824.
101. Lavelle J, Follansbee S, Trapnell CB, Buhles WC, Griffy KG,
Jung D, Dorr A, Conner J. 1996. Effect of food on the relative
bioavailability of oral ganciclovir. J Clin Pharmacol 36:238
241.
102. Jung D, Griffy K, Dorr A. 1999. Effect of food on high dose oral
ganciclovir disposition in HIV positive subjects. J Clin Pharmacol 39:161165.
103. Wilson CG, Washington N, Hardy JG, Bond SW. 1987. The
influence of food on the absorption of acyclovir: A pharmacokinetic and scintigraphic assessment. Int J Pharm 38:221
225.
104. Holdich T, Small M, Ingram J. 1988. Effect of fluid and food on
the bioavailability of 800mg acyclovir; and the bioavailability

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 12, DECEMBER 2010

of 1600 mg acyclovir. Internal report BDPS/88/002, Group


R&D, The Wellcome Foundation Ltd.
105. Welling PG, Barbhaiya RH. 1982. Influence of food and
fluid volume on chlorothiazide bioavailability: Comparison
of plasma and urinary excretion methods. J Pharm Sci 71:
3235.
106. Perez de la Cruz Moreno M, Oth M, Deferme S, Lammert
F, Tack J, Dressman J, Augustijns P. 2006. Characterization of fasted-state human intestinal fluids collected from
duodenum and jejunum. J Pharm Pharmacol 58:1079
1089.
107. Persson E, Gustafsson AS, Carlsson A, Nilsson R, Knutson L,
Forsell P, Hanisch G, Lennernas H, Abrahamsson B. 2005.
The effects of food on the dissolution of poorly soluble drugs in
human and in model small intestinal fluids. Pharm Res
22:21422151.

DOI 10.1002/jps

Potrebbero piacerti anche