Pharmaceutical Research, Vol. 25, No.

8, August 2008 ( # 2008)

DOI: 10.1007/s11095-008-9645-9

Workshop Report
Bioavailability and Bioequivalence: Focus on Physiological Factors
and Variability
Vangelis Karalis,1 Panos Macheras,1 Achiel Van Peer,2 and Vinod P. Shah3,4

Received May 6, 2008; accepted May 29, 2008; published online June 13, 2008
Abstract. This is a summary report of the EUFEPS & COST B25 conference on Bioavailability and
Bioequivalence which focused on physiological factors and variability. This conference was held at The
Royal Olympic Hotel in the centre of Athens (Greece) during the 12 of October in 2007. The issues
discussed in the conference involved physiological factors affecting drug absorption, the role of presystemic effects on bioavailability (BA), the impact of variability in bioequivalence (BE) studies, and a
final closing panel session on unresolved issues in BA/BE regulations. Several important aspects of drug
absorption were highlighted. It was presented how the complexity of gastrointestinal (GI) physiology and
the site dependent absorption can impact on drug BA. Similarly, the effects of food and formulation were
also studied. The second session focused on integrating the complexities of GI into modeling the interindividual variability of absorption and the prediction of first-pass metabolism from in-vitro data. The
necessity to measure metabolites, the value of Biopharmaceutical Classification System (BCS), and the
more recently proposed Biopharmaceutical Drug Disposition Classification System (BDDCS) were
assessed as well. This session closed with presentations of pharmacokinetic software delegates. In the
second day of the conference, the problem of high intra-subject variability in BE studies was analyzed.
Study design considerations, the use of multiple-dose studies and the role of statistics in BE were also
highlighted. Finally, the current thinking of regulatory authorities (EMEA and US-FDA) was presented.
The conference closed with a last session on unresolved issues in the regulatory level.
KEY WORDS: bioavailability and bioequivalence; highly variable drugs; impact of variability on BE
studies; physiological factors affecting drug absorption.
Contributors
Jose A. Morais, University of Lisbon, Lisbon, Portugal
Victor A. Voicu, Romanian Academy and University of Medicine and Pharmacy
Carol Davila, Bucharest
Clive G. Wilson, Strathclyde Institute for Pharmacy and Biomedical
Sciences, Glasgow, Scotland UK
Thomas Gramatte, Munich, Germany
Frank Donath, SocraTec R&D, Oberursel, Germany.
Werner Weitschies, University of Greifswald, Greifswald, Germany
Sigrid Stockbroekx, Johnson & Johnson, Beerse, Belgium
Constantin Mircioiu, University of Medicine and Pharmacy, Bucharest,
Romania Amin Rostami-Hodjegan, University of Sheffield and Simcyp Ltd,
Sheffield, UK Henning H. Blume, SocraTec R&D, Oberursel, Germany.
Geoffrey T. Tucker, University of Sheffield, Sheffield, UK
Leslie Z. Benet, University of California, San Francisco, CA, USA Meir
Bialer, Hebrew University of Jerusalem, Jerusalem
Kamal K. Midha, University of Saskatchewan and Pharmalytics Inc,
Saskatoon, Canada.
Tomas Salmonson, Medical Products Agency, Uppsala, Sweden
Barbara M. Davit, FDA, Rockville, MD, USA
George Aislaitner, National Organisation for Medicines, Athens, Greece
Alfredo Garcia-Arieta, Spanish Agency for Medicines and Health
Care Products, Madrid, Spain.
Jan Welink, Medicines Evaluation Board, The Hague, Netherlands.
1

Department of Pharmacy, University of Athens, Athens, Greece.

Johnson & Johnson, Beerse, Belgium.
3
International Pharmaceutical Federation, North Potomac, Maryland, USA.
4
To whom correspondence should be addressed. (e-mail: dr.vpshah@
comcast.net)
2

0724-8741/08/0800-1956/0 # 2008 Springer Science + Business Media, LLC

INTRODUCTION
Drug transit through the body is a composite procedure
arising from the complexity and the diversity of interactions
between the drug, physiological mechanisms and various
exogenous factors. Accordingly, the relationship between
drug intake and clinical response is considered highly
complex and is potentially affected by intrinsic and extrinsic
variables. A major cause for deviations of drugs responses
can be ascribed to product bioavailability (BA), namely the
rate and extent of drug absorption.
Various sporadic in vivo observations in the 1950s and
1960s raised the first intimations of bioequivalence (BE) with
multi-source drug products (17). It was realized that product
efficacy depends on the proportion of the drug which is
ultimately absorbed (extent of absorption) from its formulation and how rapidly the drug absorption is being held (rate
of absorption). Thus, the two key terms, extent and rate of
absorption, formed the basis of BA and BE testing. Since
then BE assessment relies on the assumption that the therapeutic effect of a drug product is a function of concentration
of the active moiety in the systemic circulation.
Both BA and BE exert a critical role in drug development
and regulatory context. Aim of this conference was: (a) to provide an insight into the physiological factors that can influence
drug absorption, (b) to address the role of pre-systemic effects
on BA, and (c) to assess the impact of variability in BE studies.

1956

1957

Bioavailability and Bioequivalence

The conference consisted of four sessions. The three first
sessions were in accordance with the objectives quoted above,
while there was also a final discussion session on unresolved
issues in BA/BE regulations.
PHYSIOLOGICAL FACTORS AFFECTING DRUG
ABSORPTION
Drug absorption is influenced by the physicochemical
properties of the drug itself, the formulation effects, and several
physiological factors. These physiological factors include gastric
emptying, intestinal motility, blood flow rate, gastrointestinal
pH, and first pass metabolism. In addition, regional differences
in GI physiology amplify the potential to affect drug absorption
processes. Changes to splanchnic blood flow and bile secretion
may lead to different first-pass metabolism of the drugs and bilesalt solubilisation of lipophilic drugs, respectively (8).
Physiological factors are subject to species variation and
hence differences in oral BA may be the result of differences in
the physiological factors mentioned above. The existence of
transporter proteins can potentially contribute to the observed
variability.
In addition, coadministration of drugs with meals or other
drugs may alter GI physiological conditions and influence drug
absorption. Aging and GI disease states may lead to alterations
in GI physiology and physiological response, resulting in further changes in the extent and rate of drug absorption.

Site Dependent Drug Absorption and Impact on Drug

Absorption
Dr. T. Gramatt (SocraTec, DE) described how site
dependent drug absorption can impact on drug absorption. It
was demonstrated that the use of various methods to
investigate regional intestinal drug absorption was shown. In
indirect approaches, absorption from the intestinal lumen is
estimated indirectly by following the appearance of the drug
within the systemic circulation. Besides, the direct approach
relies on the measurement of drug from the intestinal lumen
and can also be combined with the simultaneous measurement of plasma concentrations of drug. Intestinal perfusion
studies with several drugs (e.g. ranitidine) demonstrated that
absorption can vary depending on the site of perfusion.
It was concluded that for small hydrophilic drugs (like
paracetamol) and substances with sufficient lipophilicity there
was a uniform and efficient mucosal permeation along the
entire small intestine. However, a reduction of drug absorption from distal parts is prominent for high hydrophilic drugs.
In these cases a sharp decrease, along quite relatively short
distances of the small intestine, in AUC values was present.
According to the intestinal perfusion studies it was noted that
day-to-day variations of segmental intestinal transit times and
regional fluid fluxes can have a significant impact on net drug
absorption and contribute to high intra-subject variability.
Impact of Food on Drug Absorption

Complexity of GI Physiology and Impact on Drug

Absorption
The first session started with the presentation of Dr. C.
Wilson (Strathclyde Institute, UK) who underlined the
complexity of the GI physiology and its impact on drug
absorption. He discussed the impact of changing posture or
meal intake on the pulses of gastric emptying which in turn
leads to high variability in drug exposure. It was shown that
stomach is not homogenous and that the swollen objects or
dosage forms can alter gastric emptying
Special emphasis was placed on the application of
investigational tools such as triggered capsules or imaging
techniques (e.g. MRI) which assist in mapping the absorption
of a drug and the disintegration of formulations. Accordingly,
this can lead to the development of appropriate in vitroin
vivo correlations. It was shown that factors such as circadian
rhythm, time of dosing and meal intake control entry of drugs
into the colon and the colonic residence. The relative colon
transit of a pellet and a tablet given together was also studied
using radio-labelling techniques. Administered to a fasted
volunteer, the two preparations empty at around the same
time but in the colon the tablet is treated as a consolidated
mass and is propulsed ahead of the pellets.
In addition, the use of steady state conditions allowed
the study of drug residence in various parts of the gut. A
marked difference between the contents of right and left sides
was present for steady-state studies. Besides, this difference is
exaggerated in case of pathological conditions like active leftsided colitis. Overall, it was concluded that understanding of
gastrointestinal physiology as well as knowledge of disease
processes are necessary factors to explain variability in drug
absorption.

The presence of food in the gastrointestinal (GI) tract

can significantly alter the oral BA of drugs due to changes in
the rate and/or extent of absorption (911). These changes
can further lead to variations in efficacy and toxicity profiles
because medications are often taken under conditions of
varying food and fluid intake.
Notwithstanding the physical and chemical interactions
that may occur between drugs and specific food components,
altered postprandial absorption is generally a function the
changes associated with conversion from the fasted to the fed
state (12,13). Changes due to (a) secretion of gastric acid and
bile and pancreatic fluids, (b) modification of gastric and
intestinal motility patterns, and (c) alterations in visceral
blood and lymph flow have the most significant impact on
absorption.
The impact of food on drug absorption was addressed by
Dr. W. Weitschies (University of Greifswald). Several examples of drugs were used to demonstrate how food intake
influences GI-physiology. Using, magnetic marker monitoring
very impressive images of in vivo results were presented. For
example, it was illustrated the different position of extended
release amoxicillin-clavulanic acid tablets in the stomach
depending on whether the tablet was administered fasting,
after a first bite of a breakfast, and after the breakfast. Other
results included monitoring of drug serum concentration after
ingestion of a meal, recording of the frequency pattern of
tablet movements in the stomach and the distribution of the
capsules in the GI-tract. These techniques led to some
findings such that the storage function of the stomach, the
retention of large objects in the stomach (gastric-sieving), and
the gastro-ileocecal reflex (i.e., the existence of jet propulsion
into the colon).

1958
Formulation Effects on Drug Absorption
The formulation factors that may impact on BA and BE
can be classified into two categories (14): (a) In the first group
belong factors that can affect drug dissolution or release
which is considered as a prerequisite to the drug absorption
process. (b) The second category comprises factors related to
excipients or inactive ingredients which can influence drug
stability, absorption, and metabolism.
A presentation about formulation effects on drug absorption was given by Dr. S. Stockbroeckx (Johnson & Johnson). A
pharmaceutically stable formulation is required in order to
avoid downstream issues such as polymorphic conversions. The
beneficial properties of a drug product include enhanced solubility, stability and BA. Other issues comprise the ability to reduce odors or tastes, stabilize flavors, reduce stomach injury,
and to minimize evaporation.
The current trend in drug development is summarized in
the findings that drug candidates are obtaining higher molecular
weight, become less soluble (more lipophilic), and require more
sophisticated drug formulation technologies. These technologies can be divided into two categories. The first incorporates
the classical (or NoyesWhitney based) strategies such as: use
surfactants to alter the wetability/dispersability, increase the
surface area of the drug by reducing (micronizing) particle size
and change the salt form to increase the saturation solubility of
the drug. In the second category belong methods such as the
preparation of solid dispersions, the use of complexation and
nanosizing.
Of special interest is particle size reduction which
increases the surface area and consequently dissolution rate.
Cavitation is believed to be the main cause of size reduction.
Common methods for reducing particle size include micronization and nanonization. Micronization can be completed by
using traditional milling techniques (such as dry, wet and air
milling) as well as with other particle sizing approaches like
supercritical fluid processing. The nanosizing method usually
requires special techniques to reduce aggregation and
includes wet milling and high pressure homogenization.
ROLE OF PRE-SYSTEMIC EFFECTS
ON BIOAVAILABILITY
The second session of the conference involved presentations which described the role of pre-systemic effects on BA.
Drug absorption from the gastrointestinal tract is a very
complex and often not well characterized procedure. Both
extent and rate of drug absorption may be affected by many
factors which can be divided into three categories: The first
group comprises physicochemical factors such as pKa, aqueous
solubility, stability, diffusivity, lipophilicity, salts, surface area,
particle size and crystal form. The second category includes
physiological factors like intestinal blood flow, gastrointestinal
pH, gastric emptying, intestinal transit time, and absorption
mechanisms. The last category contains formulation factors,
namely, tablet, capsule, suspension, etc (15,16).
In order to face-off this complexity several models have been
proposed (14,1719). The first approach was pH-partition
hypothesis which was later used as a guideline for the prediction
of drug absorption. According to this hypothesis, ionizable compounds diffuse through biological membranes primarily in their

Karalis, Macheras, Peer and Shah

un-ionized forms. Hence, the degree of ionization limits the
extent of absorption of drug compounds across lipid membranes.
However, the pH-partition hypothesis is an oversimplification
which does not consider drug solubility. To face this problem,
the absorption potential (AP) concept was proposed. The AP is a
predictor of the extent of absorption and can be calculated from a
simple equation which comprises the partition coefficient,
solubility, and the dose of the drug. Some years later, the more
complex dispersion models have appeared. However, due to their
complexity these model did not exhibit a wide spread and were
followed by the much simpler mixing tank model. More recently,
the mass-balance approach was introduced which correlates
membrane permeability with the extent of drug absorption. Due
to the fact that the underlying assumption of this approach is
steady-state conditions, the rate of drug absorption cannot be
estimated. In order to face-off this demerit the Compartmental
Absorption and Transit (CAT) model has been proposed which
allows the prediction of both the extent and rate of absorption.
The Advanced Dissolution, Absorption and Metabolism (ADAM) model represents a mechanistic representation incorporating the main factors that may impact on the rate and extent of
drug absorption such as: dissolution, region-specific gut wall
permeability and metabolism, transport effects, physiological
factors (e.g., gastric emptying, intestinal transit times, distribution
of P450 enzymes in gut wall etc), effect food, and algorithms to
incorporate variability in drug absorption processes (20).
Integrating the complexities of physiology and biology of
GI tract into modeling the inter-individual variability of oral
drug absorption
The session regarding the role of pre-systemic effects on
BA started with the presentation of Dr. A. Rostami-Hodjegan
(University of Sheffield, UK). He focused on the integration of
physiology and biology encountered in GI tract into modeling
the inter-individual variability of drug absorption. Several
examples were presented which clearly demonstrated both the
progress achieved so far and the limitations that are still present.
Knowledge of the variability of the biological systems is necessary to develop useful models. Thus, it is of crucial importance to identify the variables that contribute significantly in
the process of drug absorption which will allow reliable
predictions of drug absorption. Therefore integrated gastrointestinal physiological and pharmacokinetic mechanistic models take into account variables pertinent to physicochemical/
pharmaceutical issues (e.g. disintegration, dissolution, solubility etc), physiology/pathophysiology (e.g., gastric emptying, intestinal mobility etc), fluid dynamics in the GI-tract,
fluid secretion rate and transit time in stomach and small
intestine/colon etc.
However, there are still limitations to the models developed so far. For example, a reliable prediction of the extent of
intestinal first-pass drug metabolism from in vitro data is still
challenging as the current models do not yet fully accommodate the additional complexities from gradients of enzymes
and drug transporters in the gut.
As a conclusion, it was underlined that the development of
such models should be in accordance with the general principle
that everything should be made as simple as possible, but not
simpler.
Assessment of intestinal and hepatic first-pass: Necessity to
measure metabolites? Pre-systemic elimination may occur when
orally administered drugs are subject to metabolism during their

Bioavailability and Bioequivalence

passage from the gut lumen to the systemic circulation.
Intestine and liver are the organs that may be potentially
involved. In the case of the gut and the liver, the
phenomenon is a result from the anatomical arrangement
of the splanchnic circulation which allows these organs to
act as a barrier. The role of metabolites in BE assessment
has been a controversial issue for many years (2125). One
of the major concerns raised for metabolites is the relative
variability between the measured plasma concentrations of
parent drug and metabolite.
There are several situations where metabolite data are
considered in BE studies. Firstly, metabolite data may be
used when parent drug is an inactive compound whereas the
metabolite exerts significant pharmacological activity. Another
situation for the use of metabolites arises when the serum
concentration of the parent drug is relatively low to allow a
reliable measurement. Additionally, in case of highly variable
drugs the use of metabolite data is suggested to be a possible
alternative.
The presentation of Dr. H. Blume (SocraTec, DE) focused
on the problem of assessing metabolites in BE studies. It was
highlighted the long debate which results in different regulatory
requirements in FDA and EMEA. Though, it is generally
approved that metabolite levels should be measured in BA
studies, there is no general consensus in the use of metabolites
for the assessment of BE. For bioequivalence studies determination of drug metabolites usually applies in case of non-linear
pharmacokinetics or when the metabolites are formed presystemically and contribute significantly to efficacy or safety
issues.
Prediction of Intestinal First-pass Drug Metabolism
From In-vitro Data
Intestinal first-pass effect is generally not regarded to contribute significantly to drug metabolism. However, the importance of first pass biotransformation can be clearly considered
since gut is the most important extrahepatic site of drug metabolism and comprises both phase I and II enzymes (including
many CYP enzymes).
In the presentation of Dr. G. Tucker (University of
Sheffield, UK), the objective was the prediction of intestinal
first-pass drug metabolism from in vitro data by applying
minimal models, namely models which consider the intestine as a single compartment. The models which incorporate
intestinal drug absorption can be classified into two categories: (a) the full physiologically based pharmacokinetic
models, and (b) the minimal models. The first category
comprises the segmented segregated flow model (26), the
ACAT model (27) and the ADAM model used in Simcyp
(20). The focus of was on the minimal models. Two models
were compared for a set of several CYP substrates; the wellstirred gut model and the QGut model. The simple wellstirred gut model was unable to reflect the metabolism
permeability interplay and led to poor predictions of intestinal
first-pass metabolism. However, incorporating the feature of
permeability interplay, the QGut showed an improvement in
predictability, although several sources were used to estimate
permeability clearance. Overall, a more reliable prediction
requires the application of more sophisticated models such as
ACAT and ADAM.

1959
BCS and BDDCS
The presentation of Dr. L. Benet (UCSF, USA) highlighted the value of having simple models such as the
Biopharmaceutical Classification System (BCS) and the more
recent Biopharmaceutical Drug Disposition Classification
System (BDDS) (14,28).
BCS was outlined to optimize the development of oral
dosage forms based on rate-limiting factors for absorption such
as aqueous solubility and membrane permeability. According to
BCS drugs are classified into four categories; Class I includes
drugs with high aqueous solubility and high membrane permeability, Class II comprises drugs with poor aqueous solubility
and high permeability, in Class III belong drugs with high
aqueous solubility and poor membrane permeability, while
Class IV includes drugs with poor aqueous solubility and poor
membrane permeability. The main goal of BCS was to predict in
vivo pharmacokinetic performance of drug products from
measurements of permeability and solubility and may help
decisions to obtain regulatory waivers for BE studies.
In depth examination of BCS classes revealed that compounds belonging either to Class 1 or Class 2 of BCS are eliminated primarily via metabolism, whereas the major routes of
elimination for Class 3 and Class 4 compounds are urine and bile.
The Biopharmaceutics Drug Disposition Classification System
(BDDCS) was proposed to early address issues related to drug
disposition, drug interaction and transporterenzyme interplay.
Hence, BDDCS provides a road map for the design of preclinical
and Phase 1 clinical studies without the need of running
expensive and time consuming permeability studies in humans.
In addition, it was recommended that for the definition of Class I
compounds, the regulatory agencies may use the extent of drug
metabolism (i.e., greater than 90% metabolized) instead of the
extent of drug absorption (i.e., more than 90% absorbed). A
criterion about the definition of 90% of metabolism was also
proposed which is based on the mass balance of the phase I and
phase II metabolites present in urine and feces.
IMPACT OF VARIABILITY IN BIOEQUIVALENCE
STUDIES
The third session, on the second day of the Conference,
focused on variability in BE studies. Highly variable drugs and
drug products are those exhibiting within-subject (intra-subject)
variability greater than 30% in BA parameters (29,30). This
high variability can be ascribed either to the drug substance
itself or it can be secondary to the drug product formulation.
The underlying causes of high variability include physiological,
pathological and the physicochemical properties of the drug
product. In the physiological factors belong regional pH, pancreatic or bile secretions, gastric emptying, intestinal motility,
luminal/mucosal enzymes, circadian rhythm which can significantly vary between different subjects but also within the same
subject. Other factors that can influence absorption are age,
gender, drug interactions and food intake. Of special importance, for the determination of BA and BE, is the role of firstpass effect since the activity of cytochrome enzymes show large
variability along the gastrointestinal tract. However, variability
can arise due to formulation factors of the drug product. Hence,
it is essential to keep batch-to-batch homogeneity which can be
assessed through several in vitro tests. Regardless the cause,

1960
high variability constitutes a major difficulty for the establishment of BE between the drug products.
Within-subject Variability: Design, Determination
& Demonstration
The first speaker of the second day of the conference was
Dr. K. Midha (University of Saskatchewan, Canada) who
defined within-subject variability as a measure of variability in
response within the same subject, when the subject is administered two doses of a solution on two different occasions. This
variability may be intrinsic to the drug substance and/or the
formulation. Even though, large between-subject variability may
exist for many drugs and drug products, BE is concerned with
interchangeability within a subject. It was shown that estimation
of the within-subject variability following a solution can act as the
most pure measure of variability which will further help to
understand whether the drug or the formulation is highly
variable. The use of replicate designs to measure within-subject
variability of Test and Reference formulations allows the
determination of pharmaceutical quality for each of the products.
The concept of exposure was also highlighted (31). According to exposure, the key parameters are Cmax, AUC and
AUCE which correspond to peak exposure, total exposure and
early exposure. AUCE refers to partial AUC namely the
AUC estimate truncated at the median Tmax of the reference
formulation. All these exposure measures are considered to
have clinical relevance in terms of efficacy and safety for orally
administered drug products for systemic availability. The
worthiness of AUCE for the comparison of the variability in
BE studies during the absorption phase was demonstrated using
examples where the reference formulation exhibited higher
variability than the generic product.
BE Design Considerations: Multiple Dose Versus Single Dose
Studies
Dr. A. Van Peers (Johnson & Johnson, Belgium) presentation focused on the use of multiple versus single dosing studies
for the determination of BE. Published theoretical considerations and several examples were presented. According to
existing European guidelines (32) multiple dose (steady-state)
studies would be required in the existence of non-linear
kinetics (dose- or time-dependence) and for some modified
release drug products. Other situations include high intrasubject variability or when sensitivity problems preclude the
precise estimation of plasma concentrations after single dose.
Similarly, the US authorities (33,34) propose the use of steadystate studies when there is a difference in absorption rate but
not in the extent of absorption. Other conditions involve
extended release dosage forms, very low concentrations after a
single administration of the drug and excessive variability.
A key issue in BE testing is the assessment of intra-subject
variability which can only be addressed through replicate
designs (3537). Other advantages of replicate designs include
the enrollment of a reduced number of volunteers and that
information can be retrieved about intrinsic factors which influence the formulation performance.
The presentation highlighted literature examples of BE
trials upon which the 90% confidence intervals were narrower
after multiple-dose studies than single-dose studies (38).

Karalis, Macheras, Peer and Shah

Another example demonstrated that in steady-state conditions of a modified release formulation lower intra- and
inter-subject variability were achieved. Even though, diminished variability at steady-state is favorable in terms of human
exposure in clinical trials (i.e., the required number of subjects
in the study is reduced), regulatory authorities do not support
the use of multiple dose BE studies since they may hide
differences in absorption profiles between formulations.
It was noted that at steady-state pre-dose concentrations
reflect intra-subject variability. This finding is of great importance during drug development since it offers a method to assess
the variability of a drug from early multiple-dose pharmacokinetic data.
Can Statistics Address the BE for Highly Variable Drugs?
Dr. P. Macheras (National and Kapodistrian University of
Athens, Greece) analyzed the question whether statistics can
address the BE of highly variable drugs. His presentation
started by reviewing the historical evolution of the statistical
concepts adopted in BE testing. It was highlighted the fact that
the first intimations of BE problems originated from multisource drug products and also inspired by vitamins, aspirin, and
tolbutamide (17). The measures of BE were initially defined
regarding rate and extent of absorption. Afterwards, the aim
changed to the concept of exposure which included the total,
peak, and early exposure (31). Besides, several approaches
have been proposed for the establishment of BE (33): Average
bioequivalence focuses on the comparison of means of the
product under study and the originators. In order to incorporate variability, population bioequivalence has been proposed
which assesses the total variability of the measure in the
population. Individual bioequivalence was introduced as an
effort to assess within-subject variability for test and reference
formulation, as well as the subject-by-formulation interaction.
The concept of average BE dominated along with some typical
assessment criteria such as the 80125% acceptance range, the
log-transformation of AUC and Cmax estimates, and the construction of the 90% confidence interval around the geometric
mean ratio of the two products (3234).
Reference was also made to the 75/75 rule which, at the
time it was proposed, considered as a backup to the classical
statistical tests and applied in circumstances where the power to
detect a difference was low (3941). However, due to the scientific criticism, this rule was finally discontinued (FDA 1988).
Special emphasis was given to the solutions proposed to
resolve the problem of high variability in BE assessment. Such
methods include: steady-state studies, widening of BE to prefixed constant values (e.g., 0.751.33, 0.701.43 etc.), scaled
limits, GMR-dependent scaled BE limits, and the leveling-off
scaled limits. Scaling of BE limits was initially proposed in 1995
by Boddy and his co-workers (42). Several modifications have
been proposed (4345). According to the mixed model,
unscaled limits are applied up to a switching variability value
while scaled limits are used afterwards. In another approach,
an additional regulatory criterion is imposed concomitantly
with the classic 90% CI in BE limits. This secondary criterion
suggests that the estimated ratio of geometric means should be
constrained in the range 0.801.25. Finally, the more recent
concepts of GMR-dependent scaled limits and the leveling-off
scaled BE limits have been discussed (4648). These concepts

Bioavailability and Bioequivalence

resolve the problem of discontinuity of the previous scaled
limits and offer a new alternative for reducing producer risk in
cases for highly variable drugs.
Current Thinking of EMEA Position on BE of Highly
Variable Drugs

1961
proposed in which the reference product is administered twice
and test product once to allow estimation of the intra-subject
variability of the reference product. In addition, this method
requires imposing a GMR point estimate constraint in order to
avoid the potential for declaring BE (49).
UNRESOLVED ISSUES IN BA/BE REGULATIONS

This session of the impact of variability in BE studies closed

with the presentations of EMEA and FDA representatives
which presented the current thinking of their authorities
regarding BE of highly variable drugs.
Dr. T. Salmonson (Medicinal Products Agency, Uppsala,
Sweden) shared his personal view on the current situation since
EMEA has not yet come to a formal position from the 27
European member states, although there have been several
attempts to develop a European position on this issue. The
presentation was divided into three main sections: background,
ongoing scientific debate, and what future may bring.
The first part focused on the cases in which it is now
permitted to use a wider acceptance range for the ratio of Cmax.
The answer comes from the EMEA guideline which defines
the cases in which the reference product is highly variable and
the safety/efficacy issues are justified based on PK/PD relationships or at least on clinical data.
On the ongoing scientific debate section, the questioning
of using clinical data to justify a wider acceptance range was
raised. However, it was concluded that in reality this is quite
difficult since member states disagree on level of evidence.
The methodological/philosophical aspects of scaling have also
been described. The usefulness of scaling can be justified on
the basis that is unethical to perform studies larger than
necessary and the injustice that BE may not be possible between
reference versus reference comparisons. According to Dr.
Salmonson, it is possible to define regulatory acceptance criteria
based on scaling.
Finally, it was presumed that the future will bring an update
of the current guideline along with the widening of Cmax limits
relying upon high variability and clinical justification.
Current Thinking of US-FDA on BE of Highly Variable
Drugs
The final presentation of the conference was from Dr.
Barbara Davit (Food and Drug Administration, USA) who
showed the current thinking of US-FDA on BE assessment of
highly variable drugs. Her presentation had two objectives:
the first was to discuss issues of BE submissions for new
generic drug products, while the second aim was to present a
novel approach for BE assessment.
Based on surveyed BE studies during the years 2003
2005, the drugs were categorized in respect of the encountered intra-subject variability. About 20% of the drugs with
acceptable in vivo BE studies were consistently characterized
as highly variable. On average, the acceptable BE studies of
highly variable drugs enrolled 50% more subjects than studies
of lower variable drugs.
The proposed approach for the assessment of BE is based
on simulation results. According to this approach, the appropriate sample size (in case of highly variable drugs) can be
reduced by scaling BE limits with the intra-subject variability of
the reference product. A partially replicated cross-over design is

The conference closed with a open session regarding

unresolved issues in BA/BE regulations. Several short presentations on a variety of BE issues took place. Also, an interaction
with the conference participants was offered through a panel of
three delegates of European regulatory agencies who provided
their opinions to the questions raised.
CONCLUSIONS
The variability of Highly variable drugs is due to the drug
substance pharmacokinetic characteristics, influencing the rate
and extent of drug absorption. HVDs generally exhibit withinsubject variability of >30%. The variability due to the product or
formulation factors can be easily handled through using better
formulations.
Current acceptance criteria for BE studies are that the
90% CI of the test/reference geometric mean ratio for AUC
and Cmax should fall within the limits of 0.8 to 1.25. HVD
generally do not meet these acceptance criteria unless large
number of subjects are employed in BE study. Simulation
results suggest that the number of subjects needed for BE
study of HVD can be reduced when average BE limits are
adjusted by scaling to the within-subject variability of the
reference product. Use of reference scaling is based on the
general concept that reference variability should be used as
an index for setting the standard. Reference scaling effectively decreases the sample size needed for demonstrating
BE. The within-subject variability of the reference product is
determined using partially replicated crossover study (reference product administered twice and test product administered once) design.
An additional constraint on point estimate of test/reference
geometric mean ratio (0.81.25) will eliminate the potential of a
test product with a large mean difference from the corresponding
reference product. A minimum of 24 subjects should be involved
in the BE study. The scaling factor needs to be used only when the
within-subject variability is found to be more than 30%.
REFERENCES
1. D. Melnick, M. Hochberg, and B. L. Oser. Physiological
availability of the vitamins. I. The human bioassasy technique.
J. Nutr. 30:6779 (1945).
2. A. Morrison, D. Chapman, and J. Campbell. Further studies on
the relation between in vitro disintegration time of tablets and
the urinary excretion rates of riboflavin. J. Am. Pharm. Assoc.
48:634637 (1959).
3. A. Morrison, and J. Campbell. The relationship between
physiological availability of salicylates and riboflavin and in vitro
disintegration time of enteric coated tablets. J. Am. Pharm.
Assoc. 49:473478 (1960).
4. E. Middleton, J. Davies, and A. Morrison. Relationship between
rate of dissolution, disintegration time, and physiological availability of riboflavin in sugar-coated tablets. J. Pharm. Sci.
53:13781380 (1964).

1962
5. G. Levy. Comparison of dissolution and absorption rates of
different commercial aspirin tablets. J. Pharm. Sci. 50:388392
(1961).
6. G. Levy. Effect of particle size on dissolution and gastrointestinal
absorption rates of pharmaceuticals. Am. J. Pharm. Sci. 135:78
92 (1963).
7. A. Glazko, A. Kinkel, W. Alegnani, and E. Holmes. An
evaluation of the absorption characteristics of different chloramphenicol preparations in normal human subjects. Clin. Pharmacol. Ther. 9(4):472483 (1968).
8. D. Fleisher, C. Li, Y. Zhou, L.H. Pao, and A. Karim. Drug, meal
and formulation interactions influencing drug absorption after
oral administration clinical implications. Clin. Pharmacokinet. 36
(3):233254 (1999).
9. P. G. Welling. J. Pharmacokinet. Biopharm. 5:291334 (1977).
10. L. Olanoff, T. Walle, T. Cowart, W. Walle, M. Oexmann, and E.
Conradi. Food effects on propranolol systemic and oral clearance: support for a blood flow hypothesis. Clin. Pharmacol. Ther.
40:408414 (1986).
11. S. Gupta, and L. Benet. High-fat meals increase the clearance of
cyclosporine. Pharm. Res. 7:4648 (1990).
12. J. Griffin. Drug interactions occurring during absorption from
the gastrointestinal tract. Pharmacol. Ther. 15:7988 (1981).
13. I. Walter-Sack. The influence of nutrition on the systemic
availability of drugs. Part I: drug absorption. Klin. Wochenschr.
65:927935 (1987).
14. G. Amidon, H. Lennernas, V. Shah, and J. Crison. A theoretical
basis for a biopharmaceutic drug classification: the correlation of
in vitro drug product dissolution and in vivo bioavailability.
Pharm. Res. 12:413420 (1995).
15. M. Mayersohn. Principles of drug absorption. In G. S. Banker,
and C. T. Rhodes (eds.), Modern Pharmaceutics, 2nd ed, Marcel
Dekker, New York, 1990, pp. 2390.
16. M. Gibaldi. Biopharmaceutics and clinical Pharmacokinetics, 4th
ed. Lea and Lea Febiger, Philadelphia, 1991.
17. P. Shore, B. Brodie, and C. Hogben. The gastric secretion of
drugs: a pH partition hypothesis. J. Pharmacol. Exp. Ther. 119
(3):361369 (1957).
18. N. Ho, J. Park, P. Ni, and W. Higuchi. Advanced quantitative and
mechanistic approaches in interfacing gastrointestinal drug
absorption studies in animals and humans. In W. Crouthanel,
and A. C. Sarapu (eds.), Animal models for oral drug delivery in
man: In situ and in vivo approaches, American Pharmaceutical
Association, Washington, DC, 1983, pp. 27106.
19. J. Dressman, G. Amidon, and D. Fleisher. Absorption potential:
estimating the fraction absorbed for orally administered compounds. J. Pharm. Sci. 74(5):588589 (1985).
20. Simcyp Population-based ADME simulator.
21. M. Chen, and A. Jackson. The role of metabolites in bioequivalency assessment. I. Linear pharmacokinetics without first-pass
effect. Pharm. Res. 8(1):2532 (1991).
22. M. Chen, and A. Jackson. The role of metabolites in bioequivalency assessment. II. Drugs with linear pharmacokinetics and
first-pass effect. Pharm. Res. 12(5):700708 (1995).
23. A. Jackson. The role of metabolites in bioequivalency assessment. III. Highly variable drugs with linear kinetics and first-pass
effect. Pharm. Res. 17(11):14321436 (2000).
24. A. Jackson, G. Robbie, and P. Marroum. Metabolites and
bioequivalence: past and present. Clin Pharmacokinet. 43
(10):655672 (2004).
25. K. Midha, M. Rawson, and J. Hubbard. The role of metabolites
in bioequivalence. Pharm. Res. 21(8):13311344 (2004).
26. D. Tam, R. Tirona, and K. Pang. Segmental intestinal transporters and metabolic enzymes on intestinal drug absorption.
Drug Metab. Dispos. 31(4):373383 (2003).
27. B. Agoram, W. Woltosz, and M. Bolger. Predicting the impact of
physiological and biochemical processes on oral drug bioavailability. Adv. Drug Deliv. Rev. 50:S41S67 (2001).
28. C. Wu, and L. Benet. Predicting drug disposition via application
of BCS: transport/absorption/elimination interplay and development of a biopharmaceutics drug disposition classification
system. Pharm. Res. 22(1):1123 (2005).

Karalis, Macheras, Peer and Shah

29. I. McGilveray, K. Midha, J. Skelly, S. Dighe, J. Doluisio, I.
French, A. Karim, and R. Burford. Consensus report from Bio
International 89: issues in the evaluation of bioavailability data.
J Pharm. Sci. 79(10):945946 (1990).
30. H. Blume, and K. Midha. Bio-International 92, Conference on
Bioavailability, Bioequivalence and Pharmacokinetic Studies. J.
Pharm. Sci. 82(11):11861189 (1993).
31. T. Tozer, F. Bois, W. Hauck, M. Chen, and R. Williams.
Absorption rate vs. exposure: which is more useful for bioequivalence testing. Pharm. Res. 13(3):453456 (1996).
32. Committee for Proprietary Medicinal Products (CPMP). European Agency for the Evaluation of Medicinal Products. Note for
Guidance on the Investigation of Bioavailability and Bioequivalence, London, 2001.
33. Center for Drug Evaluation and Research (CDER), Food and
Drug Administration, Statistical Approaches to Establishing
Bioequivalence. Rockville, MD, 2001.
34. Center for Drug Evaluation and Research (CDER), Food and
Drug Administration. Bioavailability and Bioequivalence Studies
for Orally Administered Drug Products. General Considerations,
Rockville, MD, 2003.
35. H. Blume, M. Elze, H. Potthast, and B. Schug. Practical
strategies and design advantages in highly variable drug studies:
multiple dose and replicate administration design. In H. H.
Blume and K. Midha (eds.), Bio-international 92: Bioavailability,
Bioequivalence, and Pharmacokinetic Studies. Medpharm, Stuttgart, 1995, pp. 117122.
36. K. Midha, M. Rawson J. Hubbard, E. Ormsby. Practical
strategies and design advantages in highly variable drug studies:
replicate design. In H. H. Blume and K. Midha (eds.), Biointernational92: Bioavailability, Bioequivalence, and Pharmacokinetic Studies. Medpharm, Stuttgart, 1995, pp. 117122.
37. V. Shah, A. Yacobi, W. Barr, L. Benet, D. Breimer, M.
Dobrinska, L. Endrenyi, W. Fairweather, W. Gillespie, M.
Gonzales, J. Hooper, A. Jackson, L. Lesko, K. Midha, P.
Noonan, R. Patnaik, and R. Williams. Evaluation of orally
administered highly variable drugs and drug formulations.
Pharm. Res. 13:15901594 (1996).
38. A. Jackson. Prediction of steady-state bioequivalence relationships using single dose data I-linear kinetics. Biopharm. Drug
Dispos. 8(5):483496 (1987).
39. Federal Register. 42:16211653 (1977).
40. Federal Register. 45:56832 (1980).
41. S. Dighe. Bioequivalence: emerging policies and procedures,
presented at Drug Information Association, Hilton Head, SC
(1985).
42. A. Boddy, F. Snikeris, R. Kringle, G. Wei, J. Oppermann, and K.
Midha. An approach for widening the bioequivalence acceptance
limits in the case of highly variable drugs. Pharm. Res. 12:1865
1868 (1995).
43. K. Midha, M. Rawson, and J. Hubbard. Bioequivalence: switchability and scaling. Eur. J. Pharm. Sci. 6:8791 (1998).
44. L. Tothfalusi, and L. Endrenyi. Limits for the scaled average
bioequivalence of highly variable drugs and drug products.
Pharm. Res. 20:382389 (2003).
45. L. Tothfalusi, L. Endrenyi, and K. Midha. Scaling or wider
bioequivalence limits for highly variable drugs and for the
special case of Cmax. Int. J. Clin. Pharmacol. Ther. 41:217225
(2003).
46. V. Karalis, P. Macheras, and M. Symillides. Geometric mean
ratio dependent scaled bioequivalence limits with levelling-off
properties. Eur. J. Pharm. Sci. 26:5461 (2005).
47. V. Karalis, M. Symillides, and P. Macheras. Novel scaled average
bioequivalence limits based on GMR and variability considerations. Pharm. Res. 21:19331942 (2004).
48. J. Kytariolos, V. Karalis, P. Macheras, and M. Symillides. Novel
scaled bioequivalence limits with levelling-off properties. Pharm.
Res. 23:26572664 (2006).
49. S. Haidar, B. Davit, M.-L. Chen, D. Conner, L. Lee, Q. Li, R.
Lionberger, F. Makhlouf, D. Patel, D. Schuirmann, and L. Yu.
Bioequivalence approaches for highly variable drugs and drug
products. Pharm. Res. 25(1):237241 (2008).