BE

5460-3 Shargel Ch10 R2 102504
10
Bioequivalence and Drug Product
Assessment, In Vivo
Dale P. Conner and Barbara M. Davit

Center for Drug Evaluation and Research,
US Food and Drug Administration, Rockville, MD, U.S.A.
1. INTRODUCTION
No topic seems so simple but stimulates such intense controversy and
misunderstanding as the topic of bioequivalence. The apparent simplicity
of comparing in vivo performance of two drug products is an illusion that
is quickly dispelled when one considers the di⁄culties and general public
misunderstanding of the accepted regulatory methodology. In the US one
often hears members of the public and medical experts alike stating various
opinions on the unacceptability of approved generic drug products based
on misconceptions regarding the determination of therapeutic equivalence
of these products to the approved reference. These misconceptions include
the belief that the US Food and Drug Administration (FDA) approves
generic products that have mean di¡erences from the reference product of
20^25% and that generic products can di¡er from each other by as much
as 45%. In addition, some incorrectly assume that, since most bioequiva-
lence testing is carried out in normal volunteers, it does not adequately
re£ect bioequivalence and therefore therapeutic equivalence in patients.
When the current bioequivalence methods and statistical criteria are
clearly understood it becomes apparent that these methods constitute a
227
Copyright © 2005 by Marcel Dekker, Inc.

5460-3 Shargel Ch10 R2 102504
228 Conner and Davit
strict and robust system that provides assurance of therapeutic equiva-

lence. In this chapter we will discuss the history, rationale and methods
utilized for the demonstration of bioequivalence for regulatory purposes
in the US as well as brie£y reviewing the bioequivalence requirements in
other countries. In addition, we will touch on some of the controversial
issues and di⁄culties in demonstrating bioequivalence for locally acting
drug products.
2. OBJECTIVES OF BIOEQUIVALENCE STUDIES

The most important concept in the understanding of bioequivalence is that
the sole objective is to measure and compare formulation performance
between two or more pharmaceutically equivalent drug products. Formula-
tion performance is de¢ned as the release of the drug substance from
the drug product leading to bioavailability of the drug substance and
eventually leading to one or more pharmacologic e¡ects, both desirable
and undesirable. If equivalent formulation performance from two products
can be established then the clinical e¡ects, within the range of normal
clinical variability, should also be equivalent. This is the same principle
that leads to an equivalent response from di¡erent lots of the brand-name
product.
When generic drugs are submitted for approval through the Abbre-
viated New Drug Application (ANDA) process in the US, they must be
both pharmaceutically equivalent and bioequivalent to be considered
therapeutically equivalent and therefore approvable. To be considered
pharmaceutically equivalent, two products must contain the same
amount of the same drug substance and be of the same dosage form with
the same indications and uses. Thus, an immediate release tablet would
not be considered pharmaceutically equivalent to an oral liquid suspen-
sion, capsule or modi¢ed release tablet. Bioequivalence means the
absence of a signi¢cant di¡erence in the rate and extent to which the
active ingredient becomes available at the site of drug action when admi-
nistered at the same molar dose under similar conditions in an appropri-
ately designed study. Two drug products are considered therapeutically
equivalent if they are pharmaceutical equivalents and if they can be
expected to have the same clinical e¡ect and safety pro¢le when adminis-
tered to patients under the conditions speci¢ed in the labeling. The FDA
believes that products classi¢ed as therapeutically equivalent can be
substituted for each other with the full expectation that the substituted
product will produce equivalent clinical e¡ects and safety pro¢le as the
original product.

5460-3 Shargel Ch10 R2 102504
Bioequivalence and Drug Product Assessment 229
3. HISTORY OF BIOEQUIVALENCE EVALUATION OF

GENERIC DRUG PRODUCTS
In 1938, the Federal Food Drug and Cosmetic Act was signed into law. The
new law required, among other things, that a‘‘new drug’’ product would need
to provide proof of safety before it could be marketed. The New Drug Appli-
cation (NDA) was established to provide a mechanism for proof of safety of
drugs to be submitted to the FDA. Regulations were promulgated as to the
form and content of the data to be submitted for an NDA. Originally only
toxicity studies were required along with informative labeling and adequate
manufacturing data. These early requirements have since evolved into the
comprehensive regulations found in Title 21 of the Code of Federal Regula-
tions Part 300, Subchapter D: Drugs for Human Use (21 CFR Part 300).
In 1962, The Kefauver ^ Harris Amendment to the Food Drug and
Cosmetic Act mandated that all new drug products subsequently approved
for marketing must have adequate evidence of e¡ectiveness, as well as safety
(1). The FDA was assigned the responsibility for receiving, reviewing, and
evaluating required data submissions, and enforcing compliance with the
law. An applicant submitting an NDA was now required to submit ‘‘substan-
tial evidence’’ in the form of ‘‘adequate and well-controlled studies’’ to
demonstrate the e¡ectiveness of the drug product under the conditions of
use described in its labeling. The new drug e¡ectiveness provision of the law
also applied retrospectively to all drugs approved between 1938 and 1962 on
the basis of safety only. The FDA contracted with the National Academy of
Sciences=National Research Council (NAS=NRC) to review this group of
drugs for e¡ectiveness. The NAS=NRC appointed 30 panels of experts and
initiated the Drug E⁄cacy Study. The panels reviewed approximately 3400
drug formulations and classi¢ed them either e¡ective or less than e¡ective
(2). The FDA reviewed the reports and any supporting data, and published
its conclusions in the Federal Register as Drug E⁄cacy Study Implementa-
tion (DESI) notices. The DESI notices contained the acceptable marketing
conditions for the class of drug products covered by this notice.
Many drug products had active ingredients and indications that were
identical or very similar to those of drug products found to be e¡ective in
the DESI review but lacked NDA’s themselves. Initially, in implementing
the DESI program, the FDA required that each of these duplicate drug pro-
ducts should have its own approved NDA before it could be legally marketed.
Later, the FDA concluded that a simpler and shorter drug application was
adequate for approving duplicate DESI drugs for marketing, and, in 1970,
created the ANDA procedure for the approval of duplicate DESI drug pro-
ducts (3^5). The FDA believed that it was not necessary for ¢rms seeking
approval of duplicate DESI drug products to establish the safety and e⁄cacy

5460-3 Shargel Ch10 R2 102504
of each new product identical in active ingredient and dosage form with a
drug product previously approved as safe and e¡ective. However, many of
the DESI notices included, as a requirement for approval of the duplicate
drug application, presentation of evidence that the ‘‘biologic availability’’of
the test product was similar to that of the innovator’s product.
Introduction in the late 1960s and early 1970s of sophisticated bioana-
lytical techniques made possible measurements of drugs and metabolites in
biological £uids at concentrations as low as a few nanograms per milliliter.
Using these methods, the disposition of drugs in the human body could be
characterized by determining pharmacokinetic pro¢les. The rate processes
of drug absorption, distribution, metabolism, and excretion could now be
quanti¢ed and related to formulation factors and pharmacodynamic e¡ects.
As these techniques were applied to investigate the relative bioavailability
of various marketed drug products, it became apparent that many generic
formulations were more bioavailable than the innovator products, while
others were less bioavailable.
In the late 1960s and early 1970s, many published studies documented
di¡erences in the bioavailability of chemically equivalent drug products,
notably chloramphenicol (6), tetracycline (7), phenylbutazone (8), and oxy-
tetracycline (9). In addition, a number of cases of therapeutic failure
occurred in patients taking digoxin. These patients required unusually high
maintenance doses and were subsequently found to have low plasma digoxin
concentrations (10). A crossover study conducted on four digoxin formula-
tions available in the same hospital at the same time revealed striking di¡er-
ences in bioavailability. The peak plasma concentrations following a single
dose varied by as much as seven-fold among the four formulations. These
¢ndings caused considerable concern because the margin of safety for
digoxin is su⁄ciently narrow that serious toxicity or even lethality can result
if the systemically available dose is as little as twice that needed to achieve
the therapeutic e¡ect.
To address this problem of bioinequivalence among duplicate drug
products, the US Congress in 1974 created a special O⁄ce of Technology
Assessment (OTA) to provide advice on scienti¢c issues, among which was
the bioequivalence of drug products. The OTA formed the Drug Bioequiva-
lence Study Panel.The basic charge to the panel was to examine the relation-
ships between chemical and therapeutic equivalence of drug products, and
to assess whether existing technological capability could assure that drug
products with the same physical and chemical composition would produce
comparable therapeutic e¡ects. Following an extensive investigation of the
issues, the panel published its ¢ndings to the US Congress in a report, dated
July 15, 1974, entitled Drug Bioequivalence (11,12). Notably, the panel con-
cluded that variations in drug bioavailability were responsible for some

5460-3 Shargel Ch10 R2 112204
instances of therapeutic failures and that analytical methodology was

available for conducting bioavailability studies in man. Several recommenda-
tions pertained to in vivo bioequivalence evaluation. The panel recom-
mended that e¡orts should be made to identify classes of drugs for which
evidence of bioequivalence is critical, that current law requiring manufac-
turers to make bioavailability information available to the FDA should be
strengthened, and that additional research aimed at improving the assess-
ment and prediction of bioequivalence was needed.
In 1977, the FDA published its Bioavailability and Bioequivalence regu-
lations under 21 CFR.The regulations were divided into Subpart AGeneral
Provisions, Subpart BProcedures for Determining the Bioavailability of
Drug Products, and Subpart CBioequivalence Requirements (13). The regu-
lations greatly aided the rational development of dosage forms of generic
drugs, as well as the subsequent evaluation of their performance. With the
publication of these regulations, a generic ¢rm could ¢le an ANDA that pro-
vided demonstration of bioequivalence to an approved drug product in lieu
of clinical trials. Subpart B de¢ned bioavailability in terms of rate and extent
of drug absorption, described procedures for determining bioavailability of
drug products, set forth requirements for submission of in vivo bioavailabil-
ity data, and provided general guidelines for the conduct of in vivo bioavail-
ability studies. Subpart C set forth requirements for marketing a drug
product subject to a bioequivalence requirement. ANDAs were generally
still restricted to duplicates of drug products approved prior to October 10,
1962 and determined to be e¡ective for at least one indication in a DESI
notice. A duplicate drug product had to meet bioequivalence requirements
if well-controlled trials showed that it was either not therapeutically equi-
valent or bioequivalent to other pharmaceutically equivalent products.
Narrow therapeutic index (NTI) drugs also had to meet bioequivalence
requirements, as did drugs with low aqueous solubility, poorly absorbed
drugs, drugs with nonlinear pharmacokinetics, drugs that underwent exten-
sive ¢rst-pass metabolism, drugs which were unstable in the GI tract, and
drugs for which absorption was limited to a speci¢c portion of the GI tract.
Finally, a duplicate drug product had to meet bioequivalence requirements
if competent medical determination concluded that a lack of bioequivalence
would have a serious adverse e¡ect in the treatment or prevention of a
serious disease or condition.
An important feature of the 1977 regulations was the provision for
waiver of in vivo bioequivalence study requirements (biowaivers) under
certain circumstances. Applicants could ¢le waiver requests for parenteral
solutions, topically applied preparations, oral dosage forms not intended to
be absorbed, gases or vapors administered by the inhalation route, and
oral solubilized dosage forms. Waivers could be granted for duplicate

5460-3 Shargel Ch10 R2 102504
DESI-e¡ective parenteral drug products (suspensions excluded) and dupli-

cate DESI-e¡ective immediate-release oral drug products which were not
on the list of FDA pharmacological classes and drugs for which in vivo
bioequivalence testing was required. Biowaivers could also be granted for
drug products in the same dosage form, but a di¡erent strength, and propor-
tionally similar in active and inactive ingredients to a drug product from the
same manufacturer for which in vivo bioavailability had been demonstrated.
Both drug products were required to meet an appropriate in vitro test
(generally dissolution) approved by the FDA.
The FDA did allow some duplicate drug versions of post-1962 drug
products to be marketed under a ‘‘paper NDA’’ policy (14).Under this policy,
in lieu of conducting their own tests, manufacturers of such duplicate drug
products could submit safety and e¡ectiveness information derived primar-
ily from published reports of well-controlled studies. However, such reports
of adequate and well-controlled studies in the literature were limited, and
the FDA sta¡ e¡ort involved in reviewing paper NDA’s became a substantial
and often ine⁄cient use of resources.
In 1984, the Drug Price Competition and Patent Term Restoration Act
(the Hatch ^ Waxman Amendments) amended the Federal Food, Drug, and
Cosmetic Act (15) by creating Section 505( j) of the Act (21 USC 355 ( j)),
which established the present ANDA approval process. Section 505( j)
extended the ANDA process to duplicate versions of post-1962 drugs, but
also required that an ANDA for any new generic drug product shall contain
information to show that the generic product is bioequivalent to the refer-
ence listed drug product. Evidence of bioequivalence was now required for
all dosage forms: tablets, capsules, suspensions, solutions, topical ointments
and creams, transdermal patches, ophthalmics, injectables, and so on. The
new law stated that a drug shall be considered to be bioequivalent to a listed
drug if ‘‘the rate and extent of absorption of the drug do not show a signi¢-
cant di¡erence from the rate and extent of absorption of the listed drug when
administered at the same molar dose and of the therapeutic ingredient under
similar experimental conditions in either a single dose or multiple doses . . . ’’
In1992,the FDA revised the Bioavailability and Bioequivalence Require-
ments of 21 CFR Part 320 to implement the Hatch ^ Waxman Amendments
(14). In its present form, 21 CFR Part 320 consists of Subpart A, General
Provisions, and Subpart B, Procedures for Determining the Bioavailability
and Bioequivalence of Drug Products. Subpart A describes general provi-
sions including de¢nitions of bioavailability and bioequivalence. Subpart B
states the basis for demonstrating in vivo bioavailability or bioequivalence
and lists types of evidence to establish bioavailability or bioequivalence,
in descending order of accuracy, sensitivity, and reproducibility. Subpart
B also provides guidelines for the conduct and design of an in vivo

5460-3 Shargel Ch10 R2 112204
bioavailability study and lists criteria for waiving evidence of in vivo bioe-
quivalence. The present biowaiver regulations now apply to solutions of all
parental preparations, including intraocular, intravenous, subcutaneous,
intramuscular, intraarterial, intrathecal, intrasternal, and interperitoneal,
but no longer permit automatic biowaivers for all topical and nonsystemi-
cally absorbed oral dosage products (16).Waivers of in vivo testing can now
be granted for ophthalmic, otic, and topical solutions. A DESI-e¡ective
immediate-release oral drug product can be granted a waiver of in vivo test-
ing, provided it is not listed in the FDA’s Approved Drug Products with Ther-
apeutic and Equivalence Evaluations as having a known or potential
bioequivalence problem (17). Other aspects of the present regulations gov-
erning biowaivers are similar to the 1977 regulations.
4. STATISTICAL EVALUATION OF BIOEQUIVALENCE DATA

Statistical evaluation of most bioequivalence studies is based on analysis of
drug blood or plasma=serum concentration data. The area under the plasma
concentration vs. time curve (AUC) is used as an index of the extent of drug
absorption. Generally, both AUC determined until the last blood sampling
time (AUC0t) and AUC extrapolated to in¢nity (AUC1) are evaluated.
Drug peak plasma concentration (Cmax) is used as an index of the rate of drug
absorption.
Criteria for approval of generic drugs have evolved since the 1970s (18).
In the early 1970s, approval was based on mean data. Mean AUC and Cmax
values for the generic product had to be within 20% of those of the brand-
name product.In addition,plasma concentration^ time pro¢les for immediate-
release products had to be reasonably superimposable. Beginning in the
late 1970s, the 75=75 (or 75=75-125) rule was added to the criteria. According
to the 75=75 rule, the test=reference ratios of AUC and Cmax had to be
within 0.75^1.25 for at least 75% of the subjects. This was an attempt to
consider individual variability in rate and extent of absorption. In the early
1980s, the power approach was applied to AUC and Cmax parameters in
conjunction with the 75=75 rule. The power approach consisted of two
statistical tests: (1) a test of the null hypothesis of no di¡erence between
formulations using the F test; and (2) the evaluation of the power to a test
to detect a 20% mean di¡erence in treatments.
Statistically, the power approach and the 75=75 rule have poor perfor-
mance, and the FDA discontinued the use of these methods in 1986. Since
1986, the FDA has used the two one-sided tests statistical procedure, also
referred to as the 90% con¢dence interval approach.The two one-sided tests
procedure encompasses two questions (19). Stated simply, the ¢rst test
asks if the test product is signi¢cantly less bioavailable than the reference

5460-3 Shargel Ch10 R2 102504
product. The second question asks if the reference product is signi¢cantly

less bioavailable than the test product. A signi¢cant di¡erence is de¢ned as
20% at the alpha equals 0.05 level. Based on these statistical criteria, the
mean test=reference ratio of the data is usually close to one. The criteria
above may be re-stated to illustrate the rationale for the 0.80^1.25 (or 80^
125%) con¢dence interval criteria. In the ¢rst case illustrated above,
test=reference ¼ 0.80 and in the second case (or bioequivalence limit)
reference=test ¼ 0.80 (expressed by convention as test=reference ¼1.25, i.e.,
the reciprocal of 0.80). This may be stated in clinical terms as follows: if a
patient is currently receiving a brand-name reference product and is
switched to a generic product, the generic product should not deliver signi¢-
cantly less drug to the patient than the brand-name product, conversely, if a
patient is currently receiving the generic product and is switched to the
brand-name reference product the brand-name product should not deliver
signi¢cantly less drug to the patient than the generic product.
The two one-sided tests procedure is the statistical procedure used to
evaluate bioequivalence rather than the hypothesis testing procedure, which
is the statistical procedure usually used to evaluate whether one treatment
produces a result signi¢cantly di¡erent from another treatment (20). This is
because, in evaluating bioequivalence studies, we want to establish whether
the di¡erence between the generic and brand-name formulation is accep-
table. Thus, the evaluation of bioequivalence data is based on the 90%
con¢dence interval approach rather than hypothesis testing (20,21). As
described above: (1) the 90% con¢dence interval encompasses the two one-
tailed tests, each carried out at the alpha ¼ 0.05 (5%) level; and, (2) the
FDA speci¢es that the bioavailability of the generic formulation relative to
the brand-name should be within 0.80^1.25 and must be known with a 90%
con¢dence. The analysis of variance (ANOVA) is applied to bioequivalence
study data to determine the 90% con¢dence limits of the di¡erences.
Until 1992, for bioequivalence statistical analysis, the FDA generally
recommended that applicants perform ANOVA on untransformed AUC
and Cmax data to determine the 90% con¢dence limits of the di¡erences. Fol-
lowing a 1991 meeting of the Generic Drugs Advisory Committee which
focused on statistical analysis of bioequivalence data, the FDA began to
recommend that applicants perform ANOVA on log-transformed data.
Since AUC and Cmax values are log-normally distributed, the Advisory Com-
mittee and FDA statisticians concluded that log-transformed data better
satisfy the assumptions underlying the ANOVA than untransformed data.
Since 1992, the FDA has formally recommended that applicants perform
ANOVA on log-transformed data and determine the 90% con¢dence inter-
val for the ratios of the test=reference least squares geometric means in
performing average bioequivalence analysis (22).

5460-3 Shargel Ch10 R2 102504
5. CURRENT METHODS AND CRITERIA FOR

DOCUMENTING BIOEQUIVALENCE
A new FDA guidance for industry posted in 2003 updates recommendations
for documentation of bioavailability and bioequivalence in regulatory sub-
missions. Bioavailability and Bioequivalence Studies for Orally Administered
Drug Products—General Considerations provides recommendations to
sponsors planning to include bioavailability and bioequivalence informa-
tion for orally administered drug products in regulatory submissions (23).
The guidance addresses how to meet the Bioavailability=Bioequivalence
Requirements set forth in 21 CFR Part 320 as they apply to oral dosage forms.
The guidance also applies to nonorally administered drug products where
reliance on systemic exposure measures is suitable to document bioavail-
ability=bioequivalence (e.g., transdermal systems, certain rectal and nasal
drug products).
There are several types of studies commonly used for demonstration of
bioequivalence. The preferred study for most orally administered dosage
forms is a two-way crossover, two-period, two-sequence single-dose study,
under fasting conditions performed in volunteers. In this design, each study
subject receives each treatment, test and reference, in random order. Plasma
or blood samples are collected for approximately three pharmacokinetic
half-lives for determination of the rate and extent of drug release from the
dosage form and absorption by each subject. A washout period is scheduled
between the two periods to allow the subjects to completely eliminate the
drug absorbed from the ¢rst dose before administration of the second dose.
Although this design is carried out for most orally absorbed drug products,
it may become impractical for drugs with long pharmacokinetic half-lives,
i.e., longer than 30 h (e.g., amiodarone, clomiphene, me£oquine). In this case
a single-dose parallel design may be used instead (24). For drugs with very
long half-lives, concentration sampling may be carried out for a period of
time corresponding to two times the median Tmax (time to Cmax) for the pro-
duct. For drugs that demonstrate low intrasubject variability in distribution
and clearance, an AUC truncated at 72 h may be used in place of AUC0t or
AUC1 (23). An alternative study design that may be useful for highly vari-
able drug products is a replicate design (23). In this design, each treatment
is repeated in the same subject on two separate occasions. This is performed
as either a partial (three-way) or full (four-way) replication of treatments.
The replicate design (22) has the advantage that fewer subjects can be used.
Because food can in£uence the bioequivalence between test and
reference products, the FDA recommends that applicants conduct bio-
equivalence studies under fasting as well as under fed conditions for
most orally administered immediate-release drug products (25). Fed

5460-3 Shargel Ch10 R2 102504
bioequivalence studies are particularly recommended for immediate-

release oral dosage forms whenever the innovator’s label makes statements
about the e¡ect of food on absorption or administration. However, if the
label states that the product should be taken only on an empty stomach, fed
bioequivalence studies are not recommended. Fed bioequivalence studies
are generally conducted using meal conditions expected to provide the
greatest e¡ects on formulation performance and gastrointestinal physiology
such that systemic drug bioavailability may be maximally e¡ected. Typically,
the drug is administered to subjects within 30 min of consuming a high-fat,
high-calorie meal. Fed bioequivalence studies should be conducted for all
modi¢ed-release oral dosage forms because the bioavailability of these pro-
ducts is likely to be altered by co-administration with meals. The FDA
recommends that these studies use a randomized, balanced, single-dose,
two-treatment (fed vs. fasting), two-period, two sequence crossover design
(25). For a few drug products, bioequivalence is evaluated only under fed
conditions because there are safety concerns associated with administration
of the product on an empty stomach.
The FDA recommends that in vivo bioequivalence studies be con-
ducted in individuals representative of the general population, taking into
account age, sex, and race factors (23). For example, if a drug product is to
be used in both sexes, the sponsor should attempt to include similar propor-
tions of males and females in the study; if the drug product is to be used pre-
dominantly in the elderly, the applicant should attempt to include as many
subjects of 60 years of age or greater as possible. Restrictions on admission
into the study should generally be based solely on safety considerations.
Bioequivalence studies should be conducted in the intended patient
population when there are signi¢cant safety concerns associated with use in
healthy subjects. For example, bioequivalence of an antineoplastic drug
intended for short-term therapy, such as etoposide, can be evaluated follow-
ing a single dose either in cancer patients in remission or in patients under
active treatment by sampling on the ¢rst day of a treatment cycle (24). For a
medication such as clozapine, on which normal subjects may experience seri-
ous orthostatic hypotension with the ¢rst dose, the most appropriate study
design is a steady-state (multiple dose) crossover bioequivalence study in
patients (24,26). These studies can be conducted either in naive patients, or
in patients who are already stabilized on the medication of interest.
6. TYPES OF EVIDENCE TO ESTABLISH BIOAVAILABILITY

AND BIOEQUIVALENCE
Subpart B of the Bioavailability and Bioequivalence Requirements in 21 CFR
Part 320 lists the following in vivo and in vitro approaches to determining

5460-3 Shargel Ch10 R2 102504
bioequivalence in descending order of accuracy, sensitivity, and reproduci-

bility (27):
in vivo measurement of active moiety or moieties in biologic £uid;

in vivo pharmacodynamic comparison;
in vivo limited clinical comparison;
in vitro comparison;
any other approach deemed appropriate by FDA.
Figure 1 illustrates, for a model of oral dosage form performance, why

the most sensitive approach is to measure the drug in biological £uids, such
as blood, plasma, or serum. The active ingredient leaves the solid dosage
FIGURE 1 The most sensitive approach in evaluating bioequivalence of two formula-

tions is to measure drug concentration in biological fluids, as illustrated in this diagram
showing the relationship between dosage form performance and therapeutic response.
Following oral dosing, the active ingredient leaves the solid dosage form, dissolves in
the gastrointestinal tract, and, following absorption through the gut wall, appears in
the systemic circulation. Formulation performance is the major factor determining the
critical steps of dosage form disintegration and drug substance dissolution prior to
absorption. All other steps following in vivo drug substance dissolution are patient- or
subject-determined processes not directly related to formulation performance. The
variability of the measured endpoint increases with each additional step in the process,
such that variability of clinical measures is quite high compared to that of blood con-
centration measures. As a result, a pharmacodynamic or clinical approach is not as
accurate, sensitive, and reproducible as an approach based on plasma concentrations.

5460-3 Shargel Ch10 R2 102504
form and dissolves in the gastrointestinal tract, and following absorption

through the gut wall, appears in the systemic circulation. The step involving
dissolution of the drug substance prior to absorption is the critical step that
is determined by the formulation. Other steps illustrated in the diagram are
patient- or subject-determined processes not directly related to formulation
performance.Variability of the measured endpoint increases with each addi-
tional step in the process. Therefore, variability of clinical measures is quite
high compared to blood concentration measures. Figure 2 shows that the
blood concentration of a drug directly re£ects the amount of drug delivered
from the dosage form.
In situations where a drug cannot be reliably measured in blood, it
may be appropriate to base bioequivalence evaluation on an in vivo test in
humans in which an acute pharmacologic (pharmacodynamic) e¡ect is
FIGURE 2 The blood concentration of a drug directly reflects the amount of drug deli-
vered from the dosage form. The corresponding responses over a wide range of doses
will be of adequate sensitivity to detect differences in bioavailability between two for-
mulations. This is illustrated for two widely different doses, D1 and D2. Any differences
in dosage form performance are reflected directly by changes in blood concentration
(R1 and R2).

5460-3 Shargel Ch10 R2 102504
measured as a function of time. Generally, the pharmacodynamic response

plotted against the logarithm of dose appears as a sigmoidal curve, as shown
in Figure 3. It is assumed that, after absorption from the site of delivery, the
drug or active metabolite is delivered to the site of activity and, through bind-
ing to a receptor or some other mechanism, elicits a quanti¢able pharma-
codynamic response. Since additional steps contribute to the observed
pharmacodynamic response, a pharmacodynamic assay is not as sensitive
to drug formulation performance as blood drug concentrations. In develop-
ing a pharmacodynamic assay for bioequivalence evaluation, it is critical to
validate the assay by selecting the correct dose. The dose should be in the
range that produces a change in response, as shown in the midportion of the
curve. In other words, the pharmacodynamic assay should be sensitive to
FIGURE 3 In evaluating bioequivalence in a study with pharmacodynamic or clinical

endpoints, it is critical to select a dose that falls on the middle ascending portion of
the sigmoidal dose^response curve. The most appropriate dose for a study based on
pharmacodynamic or clinical endpoints should be in the range that produces a change
in response (R1), as shown in the midportion of the curve (D1). A dose that is too high
will produce a minimal response at the plateau phase of the dose^response curve,
such that even large differences in dose (D2) will show little or no change in pharmaco-
dynamic or clinical effect (R2). Thus, two formulations that are quite different may
appear to be bioequivalent.

5460-3 Shargel Ch10 R2 102504
small changes in dose. A dose that is too high will produce a minimal
response at the plateau phase of the dose ^ response curve, such that even
large di¡erences in dose will show little or no change in pharmacodynamic
e¡ect. A pharmacodynamic study can be conducted in healthy subjects. The
pharmacodynamic response selected should directly re£ect dosage form
performance but may not necessarily re£ect therapeutic e⁄cacy.
If it is not possible to develop reliable bioanalytical or pharmacody-
namic assays, then it may be necessary to evaluate bioequivalence in a well-
controlled trial with clinical endpoints. This type of bioequivalence study is
conducted in patients and is based on evaluation of a therapeutic, i.e., clini-
cal, response.The clinical response follows a similar dose ^ response pattern
to the pharmacodynamic response, as shown in Figure 3. Thus, in designing
bioequivalence studies with clinical endpoints, the same considerations for
dose selection apply as for bioequivalence studies with pharmacodynamic
endpoints. As with a pharmacodynamic study, the appropriate dose for a
bioequivalence study with clinical endpoints should be on the linear rising
portion of the dose ^ response curve, since a response in this range will be
the most sensitive to changes in formulation performance. Due to high vari-
ability and the subjective nature of clinical evaluations, the clinical response
is often not as sensitive to di¡erences in drug formulation performance as a
pharmacodynamic response. For these reasons, the clinical approach is the
least accurate, sensitive, and reproducible of the in vivo approaches to
determining bioequivalence.
Most bioequivalence studies submitted to the FDA are based on mea-
suring drug concentrations in plasma. In certain cases,whole blood or serum
may be more appropriate for analysis. Measurement of only the parent drug
released from the dosage form, rather than a metabolite, is generally recom-
mended because the concentration ^ time pro¢le of the parent drug is more
sensitive to formulation performance than a metabolite, which is more
re£ective of metabolite formation, distribution, and elimination (23). Mea-
surement of a metabolite may be preferred when parent drug concentrations
are too low to permit reliable measurement. In this case, the metabolite data
are subject to a con¢dence interval approach for bioequivalence demonstra-
tion. Both the parent and metabolite are measured in cases where the meta-
bolite is formed by presystemic or ¢rst-pass metabolism and contributes
meaningfully to safety and e⁄cacy. In this case, only the parent drug data
are analyzed using the con¢dence interval approach. The metabolite data
are not subjected to con¢dence interval analysis but rather is used to provide
supportive evidence of comparable therapeutic outcome.
Urine measurements are not as sensitive as plasma measurements, but
are necessary for some drugs such as orally administered potassium chloride
(28), because serum concentrations are too low to allow for accurate

5460-3 Shargel Ch10 R2 102504
measurement of drug absorbed from the dosage form. Both cumulative

amount of drug excreted (Ae) and maximum rate of urinary excretion (Rmax)
are evaluated statistically in bioequivalence studies which rely on urine
concentrations.
The FDA accepts pharmacodynamic e¡ect methodology to approve
generic topical corticosteroid drug products (29). This approach is based
on the ability of corticosteroids to produce blanching or vasoconstriction in
the microvasculature of the skin.
Bioequivalence study designs with clinical endpoints are used for some
oral drug products that are not systemically absorbed, such as sucralfate
tablets. Bioequivalence studies with clinical endpoints generally employ a
randomized, blinded, balanced, parallel design. Studies compare the e⁄cacy
of the test product, innovator product, and placebo to determine if the two
products containing active ingredient are bioequivalent. The placebo is
included to assure that the two active treatments in the clinical trial actually
are being studied at a dose which e¡ects the therapeutic response(s). Failure
to assure that the treatments are clinically active in the trial would show that
the trial has no sensitivity to di¡erences in formulation, i.e., the response is
on the £at bottom of the dose ^ response curve (Figure 3). A generic equiva-
lent of the innovator product should be able to demonstrate bioequivalence
for selected clinical endpoint(s) that adequately re£ect drug appearance at
the site(s) of activity and therefore formulation performance. For example,
for sucralfate tablets, the clinical endpoint is duodenal ulcer healing at 4
weeks (24).The test and reference clinical responses are considered bioequi-
valent if the 90% con¢dence interval for the di¡erences in proportions
between test and reference treatment is contained within 0.20 to 0.20.
With suitable justi¢cation, bioavailability and bioequivalence may be
established by in vitro studies alone. This approach is also suitable for some
types of locally acting products, such as cholestyramine resins (30),
which form nonabsorbable complexes with bile acids in the intestine. For
cholestyramine resins, the in vitro measures of bioequivalence are based on
the rates of binding to bile acid salts. The 90% con¢dence of the test=
reference ratios of the equilibrium binding constants should fall within the
limits of 0.80^1.25.
As described above, most studies determining bioequivalence between
generic products and the corresponding brand-name products are based on
evaluation of blood concentration data in healthy subjects. It is true that
drug pharmacokinetic pro¢les may di¡er between healthy subjects and
particular types of patients. This is because some disease states a¡ect di¡er-
ent aspects of drug substance absorption, distribution, metabolism, and
elimination. However, the e¡ects of disease on relative formulation perfor-
mance, i.e., release of the drug substance from the drug product, are rare.

5460-3 Shargel Ch10 R2 102504
Bioequivalence studies are designed to measure and compare formulation

performance between two drug products within the same individuals. It is
expected that any di¡erence between in vivo drug release from the two
formulations will be the same whether the two formulations are tested in
patients or normal subjects. Thus, generic and brand-name products which
are bioequivalent can be substituted in patients because they will produce
the same e¡ect(s). This is illustrated by ¢ndings from a recent observational
cohort study comparing e¡ectiveness and safety in patients switched from
brand-name warfarin sodium tablets to generic warfarin sodium tablets
(31). The generic product was approved based on standard bioequivalence
studies in normal volunteers. The observational cohort study showed that
the two products had no di¡erence in clinical outcome measures.
7. WAIVERS OF IN VIVO BIOEQUIVALENCE BASED ON

IN VITRO DISSOLUTION TESTING
Under certain circumstances, product quality bioavailability and bioequiva-
lence can be documented using in vitro approaches (27). In vitro dissolution
testing to document bioequivalence for nonbioproblem DESI drugs remains
acceptable. In vitro dissolution characterization is encouraged for all pro-
duct formulations investigated, including prototype formulations, particu-
larly if in vivo absorption characteristics are being de¢ned for the di¡erent
product formulations. Such e¡orts may enable the establishment of an
in vitro ^ in vivo correlation.When an in vitro ^ in vivo correlation is available
(16), the in vitro test can serve as an indicator of how the product will
perform in vivo.
For immediate-release products, an in vivo bioequivalence demonstra-
tion of one or more lower strengths can be waived based on acceptable disso-
lution testing and an in vivo study on the highest strength (23). All strengths
should be proportionally similar in active and inactive ingredients. For rea-
sons of safety of study subjects, it is sometimes appropriate to conduct the
in vivo study on a strength that is not the highest. In these cases, the FDA will
consider a biowaiver request for a higher strength if elimination kinetics
are linear over the dose range, if the strengths are proportionally similar,
and if comparative dissolution testing on all strengths is acceptable. Exam-
ples of drug products for which an in vivo study is not recommended on the
highest strength due to safety include mirtazapine tablets (17) and terazosin
hydrochloride capsules and tablets (24).
For modi¢ed-release oral drug products, application of dissolution
waivers varies depending on whether the product is formulated as a beaded
capsule or tablet. For capsules in which the strength di¡ers only in the num-
ber of identical beads containing the active moiety, in vivo testing can be

5460-3 Shargel Ch10 R2 102504
waived based on acceptable dissolution testing and an acceptable in vivo

study on the highest strength. For tablets, a biowaiver may be considered
when the drug product is in the same dosage form but in a di¡erent strength,
is proportionally similar in its active and inactive ingredients, and has
the same drug release mechanism. All strengths should exhibit similar
dissolution pro¢les in at least three media (e.g., pH 1.2, 4.5, and 6.8) (23).
Applicants can request biowaivers for immediate-release products
based on an approach termed the biopharmaceutics classi¢cation system
(BCS) (32). The BCS is a framework for classifying drug substances based
on solubility and intestinal permeability. With product dissolution, these
are the three major factors governing rate and extent of absorption from
immediate-release products. The BCS classi¢es drug substances as:
Class 1: high solubility, high permeability;
Class 2: low solubility, high permeability;
Class 3: high solubility, low permeability;
Class 4: low solubility, low permeability.
The FDA believes that demonstration of in vivo bioequivalence may
not be necessary for immediate-release products containing BCS Class 1
drug substances, as long as the inactive ingredients do not signi¢cantly a¡ect
absorption of the active ingredient(s). This is because, when a drug dissolves
rapidly from the dosage form (in relation to gastric emptying) and has high
intestinal permeability, the rate and extent of its absorption are unlikely to
depend on dissolution and=or gastrointestinal transit time.
The CDER Guidance for Industry: Waiver of In Vivo Bioavailability
and Bioequivalence Studies for Immediate Release Solid Oral Dosage Forms
Based on a Biopharmaceutics Classiﬁcation System (32), recommends meth-
ods for determining drug solubility and permeability for applicants who wish
to request biowaivers based on BCS. The drug solubility class boundary is
based on the highest dose strength of the product that is the subject of the
biowaiver request. The permeability class can be determined in vivo (mass
balance, absolute bioavailability, or intestinal perfusion approaches) or
in vitro (permeation studies using excised tissues or a monolayer of cultured
epithelial cells). Dissolution testing should be conducted in three media:
0.1 N HCl or simulated gastric £uid without enzymes; pH 4.5 bu¡er, and pH
6.8 bu¡er or simulated intestinal £uid without enzymes and the f2 test
applied (23).
8. COMPLEX DRUG SUBSTANCES

There are many drug substances that may ¢t into the category of ‘‘Complex
Drug Substances’’. These include many proteins, peptides, botanicals,

5460-3 Shargel Ch10 R2 102504
synthetic hormones, biotechnology products, and complex mixtures. For

most of these drugs, the most di⁄cult problem is to demonstrate pharmaceu-
tical equivalence. In many cases, current technology is not su⁄cient to
unequivocally characterize the drug substance in two di¡erent manufac-
turer’s products or after a single manufacturer wishes to make pre- or post-
approval changes in manufacturing procedures. These de¢ciencies in drug
substance characterization methods currently may stand in the way of the
approval of generic products for many of these products containing complex
drug substances.
9. NARROW THERAPEUTIC INDEX DRUGS

There are no additional approval requirements for generic versions of NTI
drugs vs. non-NTI drugs. The FDA does not set speci¢c standards based on
therapeutic index (23,33). The bioequivalence criteria, using the 90% con¢-
dence interval approach, are quite strict; there is no need to apply stricter
criteria for NTI drugs.The current FDA position is that any generic product
may be switched with its corresponding reference listed drug.
10. FAILED BIOEQUIVALENCE STUDIES

There are several reasons why bioequivalence studies fail. Figure 4 shows
various scenarios of bioequivalence results for several hypothetical formula-
tions (labeled F1^ F7). For simplicity, the width of the 90% con¢dence inter-
val is shown as a bar, although it is important to remember that the results
are truly not an even distribution but a normal or log-normal distribution.
The log-transformed test=reference ratios from a bioequivalence study are
distributed as a bell-shaped curve, with most of the subjects’ ratios centered
around the center or mean, and fewer subjects’ ratios falling at the edges.
The top bar in Figure 4 (F1) represents a study with a 90% con¢dence inter-
val of the test to reference ratio falling between the limits of 0.80^1.25 and
the test=reference ratios centered around 1.00. This is what most applicants
would like to achieve with the to-be-marketed formulation for a given pro-
duct. The second bar (F2) also represents a 90% con¢dence interval of
test=reference ratios falling within 0.8^1.25. Although the mean test=
reference ratio is less than 1.00, the variability is very low with the result that
this product also meets the 90% con¢dence interval criteria. The remaining
bars in Figure 4 show various scenarios of failure to demonstrate bio-
equivalence. The third bar (F3) from the top depicts a situation where the
test=reference ratios are still centered around 1.00, but because of high varia-
bility and probably inadequate sample size, the 90% con¢dence interval is
very wide. This example illustrates how highly variable drugs often need

5460-3 Shargel Ch10 R2 102504
FIGURE 4 Hypothetical bioequivalence study results for formulations F1^F7 illus-

trate various scenarios of passing and failing bioequivalence criteria. The width of
each 90% confidence interval (CI) is shown as a bar, although in actuality, the
log-transformed test=reference (T=R) ratios are distributed as a bell-shaped curve.
F1 and F2 represent results of studies in which the 90% CIs of the test=reference
ratios (T=R) fall between 0.80 and 1.25 (pass bioequivalence criteria). For F1, the
ratio of T=R means (point estimate) is near 1.00. For F2, the point estimate is less
than 1.00, but because of low variability, the 90% CI of T=R ratios still falls within
acceptable limits. F3^F7 show ways in which studies fail to pass CI criteria. With
F3, the point estimate is near 1.00, but because of high variability, the 90% CI is
very wide and the drug does not pass bioequivalence criteria. F3 may pass CI cri-
teria if the number of study subjects is increased. By contrast, F4^F7 have variabil-
ity comparable to F1. F4 represents a failure on the low side (T is less bioavailable
than R), and F5 represents a failure on the high side (R is less bioavailable than T).
Since the point estimates for F3 and F4 are still within the 0.8^1.25 range, these
formulations may also meet CI criteria if a greater number of subjects are dosed.
F6 does not meet the upper bound of the 90% CI, and the point estimate exceeds
1.25. For F7, the entire CI is outside the acceptance criteria (bioinequivalence).
Formulations F6 and F7 are so different from the reference that both will still fail
CI criteria even if the number of subjects is increased.

5460-3 Shargel Ch10 R2 102504
more subjects to attain su⁄cient statistical power to pass bioequivalence

criteria. It is very likely that a new study on the same formulation would pass
if more subjects were enrolled, thereby increasing the power of the study
and decreasing the resulting width of the 90% con¢dence interval. The
next two bars (F4 and F5) show results of studies which fail to meet bio-
equivalence criteria, one a failure on the low side (test product has lower
bioavailability than reference product), the other a failure on the high side
(reference product has lower bioavailability than test product). These two
formulations have comparable variability to the formulation that passed
(F1), but fail because there is a di¡erence between the test and reference
formulations. Because the ratio of means, or point estimate, is still within
the 0.80^1.25 limits in each of these two cases, it is also possible that these
two formulations may pass another study if many more subjects were
enrolled. The product represented by the bar that is second from the bottom
(F6) not only does not meet the upper bound of the 90% con¢dence interval,
but the point estimate exceeds 1.25. It is likely that this product will not pass
even if the power of the study is increased by enrolling more subjects.
The bottom bar (F7) represents a very extreme case in which the entire
con¢dence interval is outside the acceptance criteria. In this extreme case,
the two products are bioinequivalent. The two products are so di¡erent
that it is highly improbable that repeat studies would ever demonstrate
bioequivalence.
In bioequivalence studies of generic products, the most common rea-
son for failure is that the study was underpowered with respect to the
number of subjects in the data set. The width of the con¢dence interval is
controlled by the number of subjects and by the variability of the pharma-
cokinetic measures. Studies may be underpowered for various reasons. The
applicant may have failed to enroll an adequate number of subjects. There
may be an excessive number of withdrawals, or there may be missing data
because of lost samples. Sometimes a study may fail because of outliers. For
example, noncompliant subjects may cause the study to fail. The FDA dis-
courages deletion of outlier values, particularly for nonreplicated study
designs (22).
Until recently, applicants did not include failed bioequivalence study
data in ANDA submissions. Although applicants submitting NDAs are
required to submit data from all clinical studies to the FDA, this is not the
case for ANDAs. The Food Drug and Cosmetic Act Section 505(b)(1)(A)
states that, for NDAs, all human investigations made to show whether or
not a drug is safe for use and whether such drug is e¡ective must be sub-
mitted. Similar language was not included in Section 505( j), the section
covering the submission of ANDAs. Therefore, generic ¢rms for many
years interpreted this language to mean that failed bioequivalence studies

5460-3 Shargel Ch10 R2 102504
did not have to be submitted in their ANDAs. In November of 2000, the

Advisory Committee for Pharmaceutical Science recommended that
generic applicants submit to the FDA results of all bioequivalence studies
on the to-be-marketed formulations (34). The Committee expressed the
opinion that applicants should submit the results of failed bioequivalence
studies as complete summaries, further suggesting that the FDA should do
a brief, but careful examination to identify potential problems worthy of
requesting additional information. In 2003, the FDA posted a proposed rule
relating to failed bioequivalence studies in the Federal Register for notice
and comment (35). The proposed rule would amend the Bioavailability=
Bioequivalence Regulations to require an ANDA applicant to submit data
from all bioequivalence studies that an applicant conducts on a drug product
formulation submitted for approval. The proposed rule is consistent with
the recommendations of the Advisory Committee, in that it will require
applicants to submit summary data for all bioequivalence studies other than
those upon which the applicant relies for approval.
11. BIOEQUIVALENCE EVALUATION OF GENERIC DRUG

PRODUCTS IN OTHER COUNTRIES
11.1. Canada
In Canada, applicants can seek approval of generic drug products by submit-
ting abbreviated drug submissions (ANDS) to Health Canada’sTherapeutic
Products Directorate (36). A subsequent entry (generic) product is eligible
for an ANDS against a Canadian reference product if it is given by the same
route of administration, is shown to be pharmaceutically equivalent and
bioequivalent to the Canadian reference product, and if it falls within the
same condition(s) of use as the Canadian reference product. Applicants
may also seek approval of subsequent entry products compared to a
foreign reference product if the foreign reference meets a number of speci¢c
requirements (37).
The applicant may request that the results of a single comparative
bioavailability study be extrapolated for all strengths in a product series
provided that the proportions of the medicinal and nonmedicinal ingredi-
ents fall within speci¢ed limits and exhibit the same dissolution pro¢les.
The Therapeutic Products Directorate screens ANDS upon receipt, and if
judged to be acceptable for review they are subjected to in-depth
evaluation. The Therapeutic Products Directorate has issued a number
of guidances related to comparative bioavailability studies and bio-
equivalence studies. Three of particular relevance provide guidance
on the general format of ANDS (38) and the conduct and analysis of

5460-3 Shargel Ch10 R2 102504
bioavailability=bioequivalence studies of solid oral dosage forms intended

for systemic e¡ects in immediate-release (38) and modi¢ed-release (39)
dosage forms. In general, single-dose fasting and single-dose fed bioequi-
valence studies may be designed as either two single-dose, randomized,
two-period, two-treatment crossover studies or as one single-dose, rando-
mized, four-period, four-treatment crossover study. Multiple-dose bioequi-
valence studies are also to be conducted for modi¢ed-release products
that are used at a dose likely to lead to accumulation. Bioequivalence
studies generally enroll healthy adult male subjects. Studies in patients
may be required if the use of healthy volunteers is precluded for reasons
of safety. Normally data for only the parent compound or active ingredient
is required if it can be accurately measured, e.g., measurement of the active
component may be appropriate if the parent compound is a prodrug. An
Expert Advisory Committee on Bioavailability also recommended more
stringent requirements be considered for drugs which exhibit ‘‘compli-
cated’’ pharmacokinetics, e.g., drugs with a narrow therapeutic range,
highly toxic drugs, drugs that exhibit nonlinear kinetics, etc (40). The Ther-
apeutic Products Directorate is developing guidances related to such drugs
(41,42). In general, for a drug that demonstrates ‘‘uncomplicated pharma-
cokinetics’’ in an immediate-release dosage form, the standards for bioe-
quivalence are that the 90% con¢dence interval of the relative mean
AUC0t of the test to reference product should be within 0.8^1.25 and for
Cmax only the relative mean of the test to reference product should be
between 0.8 and 1.25 (i.e., point ratio only for Cmax must fall within the lim-
its of 80^125% and no 90% con¢dence interval requirement). These stan-
dards are to be met, normally in the fasted state, on log-transformed
parameters calculated from the measured data and also data corrected for
measured drug content (percent potency of label claim). For drugs that
exhibit ‘‘complicated’’ pharmacokinetics the standards for bioequivalence
are more stringent (40). For example, for a drug that exhibits a narrow
therapeutic range the 95% con¢dence interval of the relative mean AUC0t
and Cmax of the test to reference product should be within 0.8^1.25 in sin-
gle-dose studies conducted under both fasted and fed conditions. Again,
these standards must be met on log-transformed parameters calculated
from the measured data and also from data corrected for measured drug
content (41). For drugs with nonlinear kinetics resulting in greater than
proportional increases in AUC with increasing dose, the bioequivalence
standards for ‘‘uncomplicated’’ drugs apply, i.e., 90% con¢dence limits, but
bioequivalence studies are to be conducted under fasted and fed conditions
on at least the highest strength. Similarly, for drugs with nonlinear kinetics
resulting in less than proportional increases in AUC with increasing dose,
bioequivalence studies are conducted on at least the lowest strength (42).

5460-3 Shargel Ch10 R2 112204
11.2. Japan
In Japan, the Organization for Pharmaceutical Safety and Research (OPSR)
reviews data in applications for approval of generic medications to verify
equivalency between the proposed generic and the medicine which has
already been approved. These reviews are conducted under commission
from the Japanese Ministry of Health, Labor, and Welfare (MHLW). The
Japanese National Institute of Health Sciences (NIHS) has responsibility
for preparation of guidelines and has issued a guideline for bioequivalence
studies of generic products (43). In general, applicants conduct single-dose
crossover studies using healthy adult subjects. Multiple-dose studies are
conducted for drugs that are repeatedly issued to patients. For modi¢ed-
release products, single-dose fasted and postprandial bioequivalence studies
are conducted. A high-fat diet is administered for a postprandial bioequiva-
lence study. If a high incidence of severe adverse events is indicated after
dosing in the fasted state, a fasting study is replaced with a postprandial dose
study that uses a low fat meal. If the test and reference products show a
signi¢cant di¡erence in dissolution at about pH 6.8, subjects with low gastric
acidity (achlorohydric subjects) are used unless the application of the drug
is limited to a special population. If adverse events preclude administration
to healthy subjects, then patients are used. Bioequivalence analysis is generally
based on drug concentrations in blood. Urine samples are used if there is a
rationale. The parent drug is generally measured but major active metabo-
lites may be measured instead of the parent drug, if there is a rationale. The
acceptable range of bioequivalence is generally 0.8^1.25 as the ratios of
average AUC and Cmax of the test product to reference product, when the
parameters are logarithmically distributed. For drugs with pharmacologi-
cally mild actions, a wider range can be acceptable. The acceptable ranges
for other parameters, such as Tmax, are determined for each drug. If two
products do not meet the 90% con¢dence interval criteria, they can still be
considered bioequivalent if three additional conditions are satis¢ed: (1) the
total sample size of the initial bioequivalence study is at least 20 or the
pooled sample size of initial and add-on subject studies is at least 30; (2)
the di¡erences in average values of log-transformed AUC and Cmax between
two products are between log(0.9)log(1.11); and (3) dissolution rates of test
and reference products are evaluated to be equivalent under all dissolution
testing conditions. Pharmacodynamic studies can be conducted for
products which do not produce measurable blood or urine concentrations.
In studies based on pharmacodynamic endpoints, e⁄cacy ^ time pro¢les
are compared and acceptance criteria are established by considering the
drug’s pharmacologic activity. If bioequivalence and pharmacodynamic
studies are impossible or inappropriate, a clinical study can be conducted.

5460-3 Shargel Ch10 R2 102504
The acceptance criteria in a clinical study are established by considering the

drug’s pharmacologic characteristics and activity.
11.3. European Union

The European Agency for the Evaluation of Medicinal Products (EMEA) is
in charge of supervising medicinal products for human use throughout the
European Union (EU). The EU is an international organization of Member
States that have set up common institutions to which they delegate some of
their sovereignty so that decisions on speci¢c matters of joint interest can
be made at the European level (44). The marketing of medicinal products
throughout the EU is authorized by the European Commission, which bases
its decision on recommendations made by the EMEA.
Applicants submit to the EMEA an abridged application for marketing
approval of a generic drug product that claims essential similarity to a refer-
ence product by demonstrating bioequivalence (45). The reference product
is an innovator product that has been approved on the basis of a full dossier,
including chemical, biological, pharmaceutical, pharmacological ^toxicolo-
gical and clinical data. An approved essentially similar product can be sub-
stituted for the innovator product. The EU de¢nes essential similarity as
follows: ‘‘A medicinal product is essentially similar to an original product
where it satis¢es the criteria of having the same qualitative and quantitative
composition in terms of active substances, of having the same pharmaceuti-
cal form, and of being bioequivalent unless it is apparent in the light of scien-
ti¢c knowledge that it di¡ers from the original product as regards safety and
e⁄cacy.’’ A generic ¢rm can submit to numerous EU Member States an
abridged application claiming essential similarity to a reference product
based on bioequivalence with a reference product from one Member State.
The EMEA has posted a guidance, Note for Guidance on the Investiga-
tion of Bioavailability and Bioequivalence, to provide recommendations for
study design, statistical analysis, and requirements for biowaiver requests
(45). In general, applicants conduct single-dose crossover studies in healthy
male and nonpregnant female adult subjects. Steady-state studies are con-
ducted in the case of dose- or time-dependent pharmacokinetics, for some
modi¢ed-release products. Steady-state studies may also be conducted if
drug concentrations in plasma cannot be reliably measured following a sin-
gle dose, or if high intra-individual pharmacokinetic variability precludes
the possibility of demonstrating bioequivalence in a reasonably sized sin-
gle-dose study and this variability is reduced at steady state. In most cases
bioequivalence evaluation is based on measured concentrations of the
parent compound. A metabolite can be used if concentrations of the parent
drug are too low to be accurately measured. If a metabolite contributes

5460-3 Shargel Ch10 R2 102504
signi¢cantly to activity and the pharmacokinetic system is nonlinear, parent

and metabolite are evaluated separately. For acceptance criteria, the 90%
con¢dence interval for AUC and Cmax ratios (test=reference) should fall
within 0.8^1.25. The acceptance interval may be tightened for narrow thera-
peutic range drugs. In certain cases, a wider acceptance range is used with
sound clinical justi¢cation. Statistical evaluation of Tmax is conducted when
there is a clinically relevant claim for rapid release or action or signs related
to adverse events. In this case, the nonparametric 90% con¢dence interval
for Tmax should lie within a clinically determined range. Applicants can
request exemption from in vivo bioequivalence studies based on BCS Class
1 classi¢cation, if the drug has linear pharmacokinetics, does not show evi-
dence of bioinequivalence, and does not require monitoring for critical
plasma concentrations. A bioequivalence study based on only one strength
of a product series is acceptable, provided that the choice of strength used
for the in vivo study is justi¢ed, the drug has linear pharmacokinetics, the dif-
ferent strengths are proportionally similar, and dissolution pro¢les are
acceptable. Pharmacodynamic or comparative clinical studies are con-
ducted for products intended for local use (after oral, nasal, inhalation, ocu-
lar, dermal, rectal, vaginal, etc. administration) intended to act without
systemic absorption. If there is a risk of systemic toxicity resulting from a
locally applied, locally acting medicinal product, then systemic exposure is
measured.
12. SUMMARY
Current bioequivalence methods in the US and other countries are designed
to provide assurance of therapeutic equivalence of all generic drug products
with their innovator counterparts. The sole objective of bioequivalence test-
ing is to measure and compare formulation performance between two or
more pharmaceutically equivalent drug products. For generic drugs to be
approved in the US and most other countries, they must be pharmaceutically
equivalent and bioequivalent to be considered therapeutically equivalent
and therefore approvable. In the US, a mechanism for submitting ANDA’s
for generic products was initiated in 1962 and expanded by the Hatch ^
Waxman amendment of 1984. The requirement that ANDA submissions
contain information showing that a generic drug product is bioequivalent
to the innovator product is mandated by law, under Section 505( j) of the
US Federal Food, Drug, and Cosmetic Act. Additional Federal laws,
published under Title 21 of the Code of Federal Regulations, implement
Section 505( j). Part 320 of 21 CFR, the Bioavailability and Bioequivalence
Requirements, states the basis for demonstrating in vivo bioequivalence, lists
the types of evidence to establish bioequivalence (in descending order of

5460-3 Shargel Ch10 R2 102504
accuracy, sensitivity, and reproducibility), and provides guidelines for the

conduct and design of an in vivo bioavailability study. Through the years,
the US FDA has published Guidances for Industry which address how to
meet the Bioavailability and Bioequivalence Requirements set forth in 21
CFR Part 320. The FDA makes every attempt to update these Guidances as
the need arises to ensure that they re£ect state-of-the art scienti¢c thinking
regarding the most accurate and sensitive methods available to demonstrate
bioequivalence between two products. Consulting with panels of experts
such as Advisory Committees, participating in meetings and workshops
with Academia and Industry (both in the US and abroad), and inviting public
comment on draft guidances are among the mechanisms that the FDA
employs to keep Guidance development current.
Current statistical criteria for determining acceptability of bioequiva-
lence studies in the US and in other countries assure that the test product is
not signi¢cantly less bioavailable than the reference (usually the innovator)
product, and that the reference product is not signi¢cantly less bioavailable
than the test product. The di¡erence for each of these two tests is 20%, with
the result that the test=reference ratios of the bioequivalence measures must
fall within the limits of 0.80^1.25. A generic product that does not meet these
criteria is not approved.The FDA and regulatory agencies in other countries
agree that the most accurate, sensitive, and reproducible method for deter-
mining bioequivalence is to measure drug concentrations in blood=
plasma=serum in a single-dose study using human subjects. If it is not possi-
ble to accurately and reproducibly measure drug concentrations in such bio-
logical £uids, other approaches may be used, such as measuring active
metabolite or measuring drug in urine. For locally active drug products with
little systemic availability, bioequivalence may be evaluated by pharmacody-
namic, clinical-endpoint, or highly specialized in vitro studies. Because of
the strictness of the therapeutic equivalence criteria, there is not yet a
mechanism for approving generic versions of many complex drug sub-
stances such as proteins, botanicals, and complex mixtures. Recently, the
FDA considered requesting that applicants submit results of all in vivo stu-
dies on the to-be-marketed formulation, whether these studies pass or fail
the 90% con¢dence interval criteria to identify potential problems worthy
of seeking additional information.
Biowaivers are granted in some circumstances. Conditions under
which waivers may be granted are also stipulated in 21 CFR Part 320. For
those drug products which are systemically available and have demonstrated
acceptable in vivo bioequivalence, the requirement for an in vivo study may
be waived for lower strengths only if the strengths are proportionally similar
and show acceptable in vitro dissolution by a method approved by the FDA.
Biowaivers may also be granted if an applicant satisfactorily demonstrates

5460-3 Shargel Ch10 R2 102504
that a drug dissolves rapidly from the dosage form and has high intestinal
permeability (BCS Class 1).
It should be clear that regulatory bioequivalence evaluation of generic
drug products is quite rigorous. In approving a generic product, the FDA
and the regulatory agencies of other countries make a judgement that it is
therapeutically equivalent to the corresponding reference product. A
health-care provider can substitute an approved generic product for the
brand product with assurance that the two products will produce an
equivalent therapeutic e¡ect in each patient.
ACKNOWLEDGMENTS
The authors would like to acknowledge the assistance of Mr. Donald Hare,
Ms. Rita Hassall, Dr. Henry Malinowski, Dr. Conrad Pereira, Dr. Norman
Pound, and Dr. Lizzie Sanchez in the preparation of this manuscript.
REFERENCES
1. Kefauver-Harris Amendments. Public Law (PL) 87-781, 87th Congress, Oct. 10,
1962.
2. Drug E⁄cacy Study: A Report to the Commissioner of Food and Drugs,
National Academy of Sciences, National Research Council, Washington, DC,
1969.
3. 34 Fed Regist 2673, Feb. 27, 1969.
4. 35 Fed Regist 6574, Apr. 24, 1970.
5. 35 Fed Regist 11273, Jul. 14, 1970.
6. Glazko AJ, Kinkel AW, Alegnani WC, Holmes EL. An evaluation of the absorp-
tion characteristics of di¡erent chloramphenicol preparations in normal
human subjects. Clin Pharmacol Ther 1968; 9:472^483.
7. Barr WH, Gerbracht LM, Letchen K, Plaut M, Strahl N. Assessment of the
biologic availability of tetracycline products in man. Clin Pharmacol Ther
1972; 13:97^108.
8. Chiou WL. Determination of physiological availability of commercial phenyl-
butazone preparations. Clin Pharmacol Ther 1972; 12:296^299.
9. Blair DC, Barnes RW,Wildner EL, Murray WJ. Biologic availability of oxytetra-
cycline HCl capsules: a comparison of all manufacturing sources supplying the
United States market. JAMA 1971; 215:251^254.
10. Lindenbaum J. et al. Variations in biological availability of digoxin from four
preparations. N Eng J Med 1971; 285:1344.
11. O⁄ce of Technology Assessment. Drug Bioequivalence. A Report of the O⁄ce
of Technology Assessment Drug Bioequivalence Study Panel, US Government
Printing O⁄ce,Washington, DC, 1974.
12. O⁄ce of TechnologyAssessment. Scienti¢c commentary: Drug bioequivalence.
J Pharmacokinet Biopharm 1974; 2:433^466.

5460-3 Shargel Ch10 R2 102504
13. 42 Fed Regist 1648, Jan. 7, 1977.

14. 57 Fed Regist 17950, Apr. 28, 1992.
15. Drug Price Competition and Patent Term Restoration Act of 1984, Public Law
98-417, 98 Stat 1585-1605, Sept. 24, 1984.
16. 57 Fed Regist 17998, Apr. 28, 1992.
17. Food and Drug Administration, Center for Drug Evaluation and Research.
Approved Drug Products with Therapeutic Equivalence Evaluations. 24th ed.
Superintendent of Documents. Washington, DC: US Government Printing
O⁄ce, 2004.
18. Dighe SV, Adams WP. Bioequivalence: a United States regulatory perspective.
In: Welling PG, Tse FLS, Dighe SV, eds. Pharmaceutical Bioequivalence. New
York: Marcel Dekker, Inc, 1991:347^380.
19. Schuirmann DJ. A comparison of the two one-sided tests procedure and
the power approach for assessing the equivalence of average bioavailability.
J Pharmacokinet Biopharm 1987; 15:657^680.
20. Gibaldi M, Perrier D. Absorption kinetics and bioavailability. In: Pharmacoki-
netics. 2nd New York: Marcel Dekker, 1982:145^198.
21. Hauck WW, Anderson S. A new statistical procedure for testing equivalence in
two-group comparative bioavailability studies. J Pharmacokinet Biopharm
1984; 12:83^91.
22. US Dept of Health and Human Services, Food and Drug Administration,
Center for Drug Evaluation and Research. Guidance for Industry: Statistical
Approaches to Establishing Bioequivalence. Jan. 2001.
23. US Dept of Health and Human Services, Food and Drug Administration, Center
for Drug Evaluation and Research. Guidance for Industry: Bioavailability and
Bioequivalence Studies for Orally Administered Drug ProductsGeneral
Considerations. Mar. 19, 2003.
24. Freedom of Information Sta¡, Food and Drug Administration, Center for Drug
Evaluation and Research, Rockville, MD. Summary Basis of Approval.
for Drug Evaluation and Research. Draft Guidance for Industry: Food-E¡ect
Bioavailability and Bioequivalence Studies. Dec. 2002.
26. US Dept of Health and Human Services, Food and Drug Administration, Cen-
ter for Drug Evaluation and Research. Draft Guidance for Industry: Clozapine
Tablets InVivo Bioequivalence and InVitro Dissolution Testing. Feb. 29, 2003.
27. 57 Fed Regist 29354, Jul. 1, 1992.
28. US Dept of Health and Human Services, Food and Drug Administration, Cen-
ter for Drug Evaluation and Research. Draft Guidance for Industry: Potassium
Chloride (slow-release tablets and capsules) In Vivo Bioequivalence and
InVitro Dissolution Testing. Aug. 6, 2002.
for Drug Evaluation and Research. Guidance for Industry: Topical Dermato-
logic Corticosteroids: InVivo Bioequivalence. Mar. 6, 1998.
for Drug Evaluation and Research. Interim Guidance for Industry: Cholestyra-
mine Powder InVitro Bioequivalence. Jul. 15, 1993.

5460-3 Shargel Ch10 R2 102504
31. Swenson CN, Fundak G. Observational cohort study of switching warfarin

sodium products in a managed care organization. Amer J Health Syst Pharm
2000; 57:452^455.
32. US Dept of Health and Human Services, Food and Drug Administration,Center
for Drug Evaluation and Research. Guidance for Industry: Waiver of In Vivo
Bioavailability and Bioequivalence Studies for Immediate-Release Solid Oral
Dosage Forms Based on a Biopharmaceutics Classi¢cation System. Aug. 31,
2000.
33. Nightingale S. From the Food and Drug Administration. JAMA 1998; 279:645.
34. Advisors and Consultants Sta¡, Food and Drug Administration, Center for
Drug Evaluation and Research, Rockville, MD. Meeting of the Advisory
Committee for Pharmaceutical Science. Nov. 16, 2000.
35. 68 Fed Regist 61640, October 29, 2003.
36. Health Canada. Therapeutic Products Directorate Guideline: Preparation of
Drug Submissions: Comparative Bioavailability Studies. 1997.
37. Canadian Reference product: Policy Issue from the Drugs Directorate. Use of a
non-Canadian Reference Product, Section C.08.002.1 (c) of then Food and
Drug Regulations, December 5, 1995.
38. Health Canada. Therapeutic Products Directorate Guideline: Conduct and
Analysis of Bioavailability and Bioequivalence StudiesPart A: Oral Dosage
Formulations Used for Systemic E¡ects. 1992.
39. Health Canada. Therapeutic Products Directorate Guideline: Conduct and
Analysis of Bioavailability and Bioequivalence StudiesPart B: Oral Modi¢ed
Release Formulations. 1996.
40. Health Canada. Report C. Expert Advisory Committee on Bioavailability,
Report on Bioavailability of Oral Dosage Formulations, not in Modi¢ed
Release Form, of Drugs Used for Systemic E¡ects, Having Complicated or
Variable Pharmacokinetics. 1992.
41. Health Canada. Therapeutic Products Directorate Draft Guideline: Standards
for Comparative Bioavailability Studies Involving Drugs with a Narrow
Therapeutic Range. 1997.
42. Health Canada.Therapeutic Products Directorate Draft Guideline: Bioequiva-
lence Requirements (Draft): Drugs Exhibiting Non-Linear Pharmacokinetics.
1997.
43. National Institute of Health Sciences, Japan. Guideline for Bioequivalence
Studies of Generic Products. 1997.
44. http:==www.eurunion.org=states=home.htm.
45. The European Agency for the Evaluation of Medicinal Products. Note for
Guidance on the Investigation of Bioavailability and Bioequivalence. Jul. 26,
2001.

BE

Uploaded by

Document Informationclick to expand document information

Copyright:

Available Formats

BE

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

BE

Uploaded by

Copyright:

Available Formats

5460-3 Shargel Ch10 R2 102504

Dale P. Conner and Barbara M. Davit

Copyright © 2005 by Marcel Dekker, Inc.

228 Conner and Davit

strict and robust system that provides assurance of therapeutic equiva-

2. OBJECTIVES OF BIOEQUIVALENCE STUDIES

Copyright © 2005 by Marcel Dekker, Inc.

Bioequivalence and Drug Product Assessment 229

3. HISTORY OF BIOEQUIVALENCE EVALUATION OF

Copyright © 2005 by Marcel Dekker, Inc.

230 Conner and Davit

Copyright © 2005 by Marcel Dekker, Inc.

Bioequivalence and Drug Product Assessment 231

instances of therapeutic failures and that analytical methodology was

Copyright © 2005 by Marcel Dekker, Inc.

232 Conner and Davit

DESI-e¡ective parenteral drug products (suspensions excluded) and dupli-

Copyright © 2005 by Marcel Dekker, Inc.

Bioequivalence and Drug Product Assessment 233

4. STATISTICAL EVALUATION OF BIOEQUIVALENCE DATA

Copyright © 2005 by Marcel Dekker, Inc.

234 Conner and Davit

product. The second question asks if the reference product is signi¢cantly

Copyright © 2005 by Marcel Dekker, Inc.

Bioequivalence and Drug Product Assessment 235

5. CURRENT METHODS AND CRITERIA FOR

Copyright © 2005 by Marcel Dekker, Inc.

236 Conner and Davit

bioequivalence studies are particularly recommended for immediate-

6. TYPES OF EVIDENCE TO ESTABLISH BIOAVAILABILITY

Copyright © 2005 by Marcel Dekker, Inc.

Bioequivalence and Drug Product Assessment 237

bioequivalence in descending order of accuracy, sensitivity, and reproduci-

in vivo measurement of active moiety or moieties in biologic £uid;

Figure 1 illustrates, for a model of oral dosage form performance, why

FIGURE 1 The most sensitive approach in evaluating bioequivalence of two formula-

Copyright © 2005 by Marcel Dekker, Inc.

238 Conner and Davit

form and dissolves in the gastrointestinal tract, and following absorption

Copyright © 2005 by Marcel Dekker, Inc.

Bioequivalence and Drug Product Assessment 239

measured as a function of time. Generally, the pharmacodynamic response

FIGURE 3 In evaluating bioequivalence in a study with pharmacodynamic or clinical

Copyright © 2005 by Marcel Dekker, Inc.

240 Conner and Davit

Copyright © 2005 by Marcel Dekker, Inc.

Bioequivalence and Drug Product Assessment 241

measurement of drug absorbed from the dosage form. Both cumulative

Copyright © 2005 by Marcel Dekker, Inc.

242 Conner and Davit

Bioequivalence studies are designed to measure and compare formulation

7. WAIVERS OF IN VIVO BIOEQUIVALENCE BASED ON

Copyright © 2005 by Marcel Dekker, Inc.

Bioequivalence and Drug Product Assessment 243

waived based on acceptable dissolution testing and an acceptable in vivo

8. COMPLEX DRUG SUBSTANCES

Copyright © 2005 by Marcel Dekker, Inc.

244 Conner and Davit

synthetic hormones, biotechnology products, and complex mixtures. For