[Federal Register Volume 61, Number 102 (Friday, May 24, 1996)]
[Notices]
[Pages 26182-26186]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 96-13106]



-----------------------------------------------------------------------

DEPARTMENT OF HEALTH AND HUMAN SERVICES
Food and Drug Administration
[Docket No. 94D-0401]


Bioequivalence Guidance, 1996; Availability

AGENCY: Food and Drug Administration, HHS.

ACTION: Notice.

-----------------------------------------------------------------------

SUMMARY: The Food and Drug Administration (FDA) is announcing the 
availability of the revised guidance document entitled ``Bioequivalence 
Guidance, 1996'' prepared by the Center for Veterinary Medicine (CVM). 
The availability of a draft guideline entitled ``Bioequivalence 
Guideline (Draft) 1994'' was announced in the Federal Register of March 
1, 1995 (60 FR 11097) (hereinafter referred to as the 1994 draft 
guideline). The 1994 draft guideline was a revision of the 1990 version 
and covered the following areas: General considerations, blood level 
studies, pharmacologic endpoints, clinical endpoints, and human food 
safety. The guidance is intended to assist sponsors of new animal drug 
applications (NADA's) in the design and analysis of in vivo 
bioequivalence studies. This notice addresses comments submitted on the 
1994 draft guideline.

DATES: Written comments on the guidance document may be submitted at 
any time.

ADDRESSES: Submit written requests for single copies of the guidance 
document entitled ``Bioequivalence Guidance, 1996'' to the 
Communications and Education Branch (HFV-12), Center for Veterinary 
Medicine, Food and Drug Administration, 7500 Standish Pl., Rockville, 
MD 20855, 301-594-1755. Send two self-addressed adhesive labels to 
assist that office in processing your requests. Submit written comments 
on the guidance document to the Dockets Management Branch (HFA-305), 
Food

[[Page 26183]]

and Drug Administration, 12420 Parklawn Dr., rm. 1-23, Rockville, MD 
20857. Requests and comments should be identified with the docket 
number found in brackets in the heading of this document. A copy of the 
guidance document and received comments may be seen at the Dockets 
Management Branch between 9 a.m. and 4 p.m., Monday through Friday.

FOR FURTHER INFORMATION CONTACT: Melanie R. Berson, Center for 
Veterinary Medicine (HFV-135), Food and Drug Administration, 7500 
Standish Pl., Rockville, MD 20855, 301-594-1643.
SUPPLEMENTARY INFORMATION: FDA is announcing the availability of the 
revised guidance entitled ``Bioequivalence Guidance, 1996''. The 
guidance may be used by sponsors of NADA's for the design and analysis 
of in vivo bioequivalence studies.
    In a notice published in the Federal Register of March 1, 1995 (60 
FR 11097), FDA announced the availability of the 1994 draft guideline 
entitled ``Bioequivalence Guideline (Draft) 1994''. The 1994 draft 
guideline was based on an April 1990 bioequivalence guidance and 
reports from panel presentations at the 1993 Veterinary Drug 
Bioequivalence Workshop held in Rockville, MD. New topics addressed in 
the 1994 draft guideline included: Bioequivalence overdose studies, 
testing for multiple strength solid oral dosage forms, assay 
considerations, area under the curve and maximum blood concentration as 
pivotal parameters, and blood level studies with good laboratory 
practice tissue residue depletion studies for generic products for food 
animals. Interested persons were given until May 30, 1995, to comment 
on the 1994 draft guideline.
    Comments on the 1994 draft guideline were received from a 
pharmaceutical company and an industry group. The 1994 draft guideline 
has been revised as a result of these comments and from internal 
discussions within CVM. In the following section on received comments 
and CVM responses, the page numbers and sections refer to those found 
in the 1994 draft guideline.
    1. Section II.E. Dose Selection. The comment objected to the use of 
the term ``overdose bioequivalence study'' since ``overdose'' has 
toxicological connotations.
    CVM accepts the comments and will change the wording from 
``overdose'' to ``higher than approved dose.''
    2. Section II.F. Multiple Strengths of Solid Oral Dosage Forms. One 
comment asked for the rationale for requiring two bioequivalence 
studies in order to obtain approval when there are more than three 
strengths of exactly proportional formulations.
    CVM accepts the comment and has modified the guidance to allow more 
flexibility in the determination of the need for more than one 
bioequivalence study for multiple strengths of solid oral dosage forms. 
The guidance has been modified to read as follows:
    The generic sponsor should discuss with CVM the appropriate in 
vivo bioequivalence testing and in vitro dissolution testing to 
obtain approval for multiple strengths (or concentrations) of solid 
oral dosage forms.
    CVM will consider the ratio of active to inactive ingredients 
and the in vitro dissolution profiles of the different strengths, 
the water solubility of the drug, and the range of strengths for 
which approval is sought.
    One in vivo bioequivalence study with the highest strength 
product may suffice, if the multiple strength products have the same 
ratio of active to inactive ingredients and are otherwise identical 
in formulation.
    In vitro dissolution testing should be conducted, using an FDA 
approved method, to compare each strength of the generic product to 
the corresponding strength of the reference product.
    3. Section II.G. Manufacturing of Pilot Batch (``Biobatch''). One 
comment requested that terms such as `pilot' and `biobatch' need to be 
precisely defined in this document or reference made to the 
manufacturing guidelines.
    CVM refers the reader to CVM's ``Animal Drug Manufacturing 
Guidelines, 1994'' for definition of terms.
    4. Section III.A. Assay Considerations. One comment requests that 
CVM should adopt the same guidance as established in the joint 
industry/academia conference on ``Analytical Methods Validation: 
Bioavailability, Bioequivalence, and Pharmacokinetics Studies'' 
published in several journals including the Journal of Pharmaceutical 
Sciences, 81(3), 309-312, 1992.
    CVM does not agree with this comment. The substance of CVM's 
guidance does not differ substantially from those used by CDER. Any 
difference is the result of CVM's interest in maintaining consistency 
among its analytical criteria for drug residues in the edible tissues. 
Drug residue measurement in edible tissues is specific to animal drugs 
and is not applicable to CDER (human drugs).
    5. Section III.C.6.a. Area Under the Curve (AUC) Estimates. One 
comment questioned whether AUC by the linear trapezoidal rule is the 
preferred method to estimate AUC, and noted that the method is subject 
to substantial error when data points are widely spaced (e.g., during 
the terminal exponential disposition phase).
    CVM accepts the comment and will modify the wording in the guidance 
to acknowledge that methods other than the linear trapezoidal rule may 
be used for estimating AUC, but the alternative method should be 
accompanied by appropriate references.
    6. Section III.C.6.a. One comment questioned the reason to equate 
AUC over a dosing interval at steady-state to single-dose AUC zero to 
infinity. The comment stated that this relationship only holds if 
pharmacokinetics are linear over the relevant dose range and one of the 
prime reasons for doing a multiple-dose bioequivalence study is when 
kinetics are nonlinear.
    CVM has modified the guidance to read as follows:
    Under steady state conditions, AUC0-t equals the full 
extent of bioavailability of the individual dose (AUC0-INF), 
assuming linear kinetics. For drugs which are known to follow 
nonlinear kinetics, the sponsor should consult with CVM to determine 
the appropriate parameters for the bioequivalence determination.
    7. Section III.C.6.c. Determination of Product Bioequivalence. One 
comment requested that the sponsor should be allowed to extend the 
range of acceptable bioequivalence limits for drugs exhibiting highly 
variable pharmacokinetics, if adequate justification is provided.
    CVM accepts the comment and has modified the guidance to include 
the following statement:
    The sponsor and CVM should agree to the acceptable bounds for 
the confidence limits for the particular drug and formulation during 
protocol development. If studies or literature demonstrate that the 
pioneer drug product exhibits highly variable kinetics, then the 
generic drug sponsor may propose alternatives to the generally 
acceptable bounds for the confidence limits.
    8. One comment requested that the repeated references to flip-flop 
kinetics should be replaced by the more general term ``prolonged 
absorption.''
    CVM accepts the comment and has replaced the term ``flip-flop 
kinetics'' with ``sustained or prolonged absorption.''
    9. One comment requested that the Bioequivalence Guidance provide 
more detail on evaluation of Production Drugs and Short Term 
Therapeutic Treatments in Feed (Staff Manual Guide 1240.4145).
    CVM does not agree with the request to elaborate on combination 
drugs for use in feed. The focus of the Bioequivalence Guidance is the 
approval of generic animal drugs, although many of the principles may 
be applied to blood level studies conducted for other purposes. CVM 
considers it beyond the scope and intent of this guidance to discuss 
combination approvals for feeds.
    10. Page 1, section I. INTRODUCTION, fifth paragraph. One

[[Page 26184]]

comment requested insertion of the following paragraph:
    Tissue residue studies will not normally be required if blood 
concentration curve shape and depletion time through the reference 
product's withdrawal time are the same for generic and reference 
products. Tissue residue studies will normally be required where the 
blood levels cannot be measured prior to the elapse of the reference 
product's withdrawal period.
    CVM does not agree with the change proposed by this comment. The 
pivotal parameters for the drug concentration versus time curve are AUC 
and CMAX. CVM does not intend to evaluate curve shape and 
depletion time as pivotal parameters. For clarity, however, the 
guidance has been modified to read as follows:
    The Center has concluded that the tissue residue depletion of 
the generic product is not adequately addressed through 
bioequivalence studies. Therefore, ANADA's for drug products for 
food-producing animals will generally be required to include 
bioequivalence and tissue residue studies. A tissue residue study 
will generally be required to accompany clinical end-point and 
pharmacologic end-point bioequivalence studies, and blood level 
bioequivalence studies that can not quantify the concentration of 
the drug in blood throughout the established withdrawal period.
    11. Page 2, section II.A. Selection of Reference Product for 
Bioequivalence Testing, second paragraph. One comment suggested that 
the paragraph should read ``but remains eligible to be copied, then the 
first approved and available generic copy of the pioneer should be used 
* * *.''
    CVM accepts the comment and has reworded the paragraph.
    12. Page 5, first full paragraph. One comment suggests that 
multiple bioequivalence studies at different doses should only be 
required if the pharmacokinetics are not linear.
    CVM accepts the recommendation and has modified the guidance to 
read as follows:
    For products labeled for multiple claims involving different 
pharmacologic actions at a broad dose range (e.g., therapeutic and 
production claims), a single bioequivalence study at the highest 
approved dose will usually be adequate. However, multiple 
bioequivalence studies at different doses may be needed if the drug 
is known to follow nonlinear kinetics. The sponsor should consult 
with CVM to discuss the bioequivalence study or studies appropriate 
to a particular drug.
    13. Page 6, section III.A.1. Concentration Range and Linearity. One 
comment proposed that ``at least 5-8 concentrations'' is vague and 
suggested ``at least 5 concentrations.''
    CVM accepts the comment and has changed the wording to ``at least 5 
concentrations.''
    14. Page 7, section III.A.4. Specificity. One comment requested 
that CVM provide further detail on statistical methods for 
demonstrating ``parallelism and superimposability.'' Analysis of 
variance is used to compare means but could be used to compare slopes 
in this case. This is computationally straightforward for linear curves 
but nonlinear curves (e.g., microbiological assays) pose unique 
problems.
    CVM's response is that the type of statistical procedure used to 
process data demonstrating parallelism and superimposability of curves 
depends on the nature of the experimental data. CVM is allowing the 
sponsor the flexibility to determine the algorithm used to evaluate 
data. Whatever statistical procedure is used should be justified by the 
sponsor.
    The use of microbiological assays for drug analysis will be 
addressed in a future CVM guidance.
    15. Page 8, sections III.A.5. Accuracy (Recovery) and III.A.6. 
Precision. One comment requested that ``replicate injections'' be 
changed to ``replicates.''
    CVM accepts the comment.
    16. Page 8, section III.A.6. Precision. One comment stated that the 
suggested coefficient of variation of  10 percent for 
concentrations at or above 0.1 micrograms per milliliter (mL) is too 
stringent. The comment suggested 15 percent as an 
alternative coefficient of variation to target.
    CVM does not agree with this comment. In light of today's 
analytical technology, 10 percent coefficient of variation 
is not unreasonable and is consistent with CVM policy in other 
analytical areas. In addition, CVM does not believe anything is gained 
by a detailed analysis of the sources of variation in analytical 
results.
    17. Page 8, section III.A.7. Analyte Stability, second paragraph. 
One comment recommended that stability samples at only two 
concentrations are necessary, rather than three as suggested in the 
1994 draft guideline. It is critically important to validate the assay 
before conduct of the bioequivalence study. However, analyte stability 
cannot be done without the use of more animals than required by the 
bioequivalence study so as to have a valid method in place prior to 
study initiation. It is impossible to store and begin analyzing 
stability samples throughout the duration of the bioequivalence study 
analysis phase unless the method has been validated prior to that 
study's initiation.
    CVM does not agree with this comment. No study should be undertaken 
until the analytical methods that will be used to develop the data are 
properly validated and shown to be operating in a state of control in 
the laboratory. This means that after the method is validated, the 
laboratory intending to use the method for a study, must practice with 
the method to assure full familiarization with technical details. CVM 
does not make any recommendation on how much practice is required. This 
depends on the complexity of the method and on the experience of the 
laboratory.
    18. Page 8, section III.A.8. Analytical System Stability. One 
comment stated that it was unclear how the use of standards (of 
multiple concentrations) repetitively run to assure analytical system 
stability differs from quality control methods of assuring the same 
thing.
    CVM accepts the comment that the wording on the use of standards 
may be unclear. The guidance section on ``Assay Considerations'' has 
been extensively reworded for clarity.
    19. Page 9, section III.B.1. Dosing by Labeled Concentration. One 
comment asked how the assay prior to study will be used to ensure 
specifications. What actions can the sponsor take if the pioneer assays 
at -5 percent while the generic assays at +5 percent.
    CVM's response is that the pioneer and generic products should be 
assayed to determine that the particular lots are within 
specifications. No action can be taken if the pioneer assays at -5 
percent while the generic assays at +5 percent.
    For clarity, the guidance has been reworded to read as follows: 
``To maximize the ability to demonstrate bioequivalence, the Center 
recommends that the potency of the pioneer and generic lots should 
differ by no more than 5% for dosage form products.''
    20. Page 10, section III.B.2. Single Dose vs Multiple Dose Studies. 
One comment questioned whether documentation of flip-flop kinetics is 
necessary.
    CVM agrees with this comment and has modified the guidance to read 
as follows:
    A multiple dose study may also be needed when assay sensitivity 
is inadequate to permit drug quantitation out to 3 terminal 
elimination half-lives beyond the time when maximum blood 
concentrations (Cmax) are achieved, or in cases where prolonged 
or delayed absorption2 exist. The determination of prolonged or 
delayed absorption (i.e., flip-flop kinetics) may be made from pilot 
data, from the literature, or from the CVM database on the 
particular drug or family of drugs.
    21. Page 11, section III.B.4. Fed vs Fasted State, last paragraph. 
One comment stated that it was unclear whether studies in both the fed 
and fasted states should be required for enteric-coated or sustained 
release oral

[[Page 26185]]

products. If the referenced product is limited to administration either 
in the fed or the fasted state, then the test formulation should also 
be administered in the same situation conforming to the reference 
product's label.
    CVM agrees with this comment and has modified the guidance as 
follows:
    If a pioneer product label indicates that the product is limited 
to administration either in the fed or fasted state, then the 
bioequivalence study should be conducted accordingly. If the 
bioequivalence study parameters pass the agreed upon confidence 
intervals, then the single study is acceptable as the basis for 
approval of the generic product.
    However, for certain product classifications or drug entities, 
such as enteric coated and oral sustained release products, 
demonstration of bioequivalence in both the fasted and fed states 
may be necessary, if the drug is highly variable under feeding 
conditions, as determined from the literature or from pilot data. A 
bioequivalence study conducted under fasted conditions may be 
necessary to pass the confidence intervals. A second smaller study 
may be necessary to examine meal effects. CVM will evaluate the 
smaller study with respect to the means of the pivotal parameters 
(AUC, CMAX). The sponsor should consult with CVM prior to 
conducting the studies.
    22. Page 12, section III.C.2. Protein Binding. One comment stated 
that it is not clear from the 1994 draft guideline to what extent the 
protein binding must be nonlinear within the therapeutic dosing range, 
nor how determination of linearity is to be conducted. If it is a 
judgment and not a statistical criterion, then the parameters within 
which that judgment is made need to be determined prior to embarking 
upon the abbreviated NADA. In addition, the type of blood protein to 
which the drug binds is only pertinent in very unique situations (i.e., 
low capacity protein binding situations). These determinations of the 
type of blood protein to which the drug binds are very tedious, time-
consuming and expensive technical studies that may only rarely be 
relevant, whereas the magnitude of protein binding is critical. The 
type of blood protein to which the drug binds is only a consideration 
if prior data indicate it is a concern. There are numerous instances 
where CVM requires additional studies ``if ------------------ is known 
to occur.'' What are the criteria for knowing? This general statement 
could lead to intractable situations. Specifically for this section, 
the wording allows CVM to require protein binding studies for all 
approvals. A proposal would be to first evaluate the blood profiles 
observed in the pilot studies to see if there is evidence of such 
binding (multicompartment phenomena). If not, then eliminate the need 
for further studies. For combination approvals, the necessary 
fractionation and assessment of matrix effects using micro methods 
would be a formidable task.
    CVM notes that the Bioequivalence Guidance is not intended to 
address combination drug approvals. The issue of protein binding for 
generic approvals would be addressed only if literature or pilot data 
indicate that protein binding is significant to the drug in question. 
For clarity, however, the guidance has been modified to read as 
follows:
    However, if nonlinear protein binding is known to occur within 
the therapeutic dosing range (as determined from literature or pilot 
data), then sponsors may need to submit data on both the free and 
total drug concentrations for the generic and pioneer products.
    23. Page 14, section III.C.4. Cross-over and Parallel Design 
Considerations, last sentence. One comment proposed that the pilot data 
be used in support of alternative study designs during discussions with 
CVM.
    CVM agrees with the comment. The guidance statement has been 
modified to read as follows: ``The use of alternative study designs 
should be discussed with CVM prior to conducting the bioequivalence 
study. Pilot data or literature may be used in support of alternative 
study designs.''
    24. Page 15, top paragraph. One comment regarding the duration of 
washout time was that prolonged tissue binding may not be a consequence 
if drug concentrations in plasma are less than the limit of detection. 
The onus is on the sponsor for having a sufficiently long washout 
period to allow the second period of the cross-over study to be 
applicable in the statistical analysis. If sequence effects are noted, 
it must be emphasized that at the very minimum the same data from the 
first period alone can be evaluated as a parallel design study.
    CVM agrees with the comments and has modified the paragraph in the 
guidance to read as follows:
    The washout period should be sufficiently long to allow the 
second period of the cross-over study to be applicable in the 
statistical analysis. However, if sequence effects are noted, the 
data from the first period may be evaluated as a parallel design 
study.
    25. Page 15, section III.C.6.a., AUC Estimate. One comment stated 
that it is implied from the discussions regarding AUC and CMAX 
that ratio testing (the ratio of the test versus the reference product) 
is considered to be the more appropriate comparison rather than the 
difference between the test and the reference product. This is not 
universally accepted as the case. The responsibility for whether the 
difference between the two is used or the ratio of the two is used 
should be placed upon the sponsor and should be concurred with by CVM 
prior to conduct of the study.
    CVM does not agree with nor completely understand the comment's 
interpretation of the guidance. CVM has, however, changed the word 
``ratio'' to ``comparison'' in the following sentence:
    The comparison of the test and reference product value for this 
noninfinity estimate provides the closest approximation of the 
measure of uncertainty (variance) and the relative bioavailability 
estimate associated with AUC0-INF' the full extent of product 
bioavailability.
    26. Page 15, section III.C.6.a. One comment stated that AUC0-
INF is an estimated value and questioned how CVM intends this to be 
derived using ``model independent methods?''
    CVM has added the following statement to the guidance: ``The method 
for estimating the terminal elimination phase should be described in 
the protocol and the final study report.''
    27. Page 16, section III.C.6.b. Rate of Absorption. One comment 
requested that the revised guidance define CMIN. The 1994 draft 
guideline stated that three successive CMIN values should be 
provided. The comment proposes that to determine a steady state 
concentration, the values should be regressed over time and the 
resultant slope should be tested as being different from zero.
    CVM agrees with the comment and has modified the guidance to read 
as follows:
    When conducting a steady-state investigation, data on the 
minimum drug concentrations (trough values) observed during a single 
dosing interval (CMIN) should also be collected. Generally, 
three successive CMIN values should be provided to verify that 
steady-state conditions have been achieved. Although CMIN most 
frequently occurs immediately prior to the next successive dose, 
situations do occur with CMIN observed subsequent to dosing. To 
determine a steady state concentration, the CMIN values should 
be regressed over time and the resultant slope should be tested for 
its difference from zero.
    28. Page 16, section III.C.6.c. Determination of Product 
Bioequivalence. One comment states that for multiple dose studies, 
CMAX and AUCO-t are applicable only if done at steady state. 
It is not clear from the current description that these must be steady 
state values to have the appropriate interpretation for bioequivalence 
testing.
    CVM does not agree with the comment because a multiple dose 
bioequivalence study could be conducted with a drug that never achieves 
steady-state. However, the pioneer and generic products CMAX and

[[Page 26186]]

AUCO-t should be equivalent at any dosing interval whether or not 
steady-state is achieved.
    29. Page 17, section III.D. Statistical Analysis, second paragraph. 
The choice of whether to use untransformed data should be made by the 
sponsor based on whether transformation is necessary to allow for 
homogeneity of variance. It should not be determined prior to the study 
because the data should dictate which transformation, if any, is 
required.
    CVM does not agree with this recommendation. The sponsor has the 
option to use untransformed or log transformed data, but the decision 
should be made prior to conducting the study.
    30. Page 19, section III.D., second from the last paragraph 
relating to selection of confidence interval. One comment noted that 
CVM states that in general the confidence interval for untransformed 
data should be 80 to 120. Firstly, percent should be specified. 
Secondly, emphasis should be added that these are general rather than 
the adamant and steadfast specifications of CVM. The opinion of many 
statisticians with considerable experience in this field is that the 
20 percent interval is entirely too restrictive. In the 
animal health market, the potential cost to evaluate generics or 
combinations may be so great as to preclude bringing a useful drug/
combination to the market.
    CVM has made the requested editorial changes. However, CVM will 
continue to accept 20 percent as the acceptable confidence 
interval for the pivotal parameters. CVM invites sponsors to submit 
data to justify broadening the confidence interval for a particular 
drug.
    31. Page 20, section IV.B. Statistical Analysis. One comment noted 
that for pharmacologic endpoint studies as described, it appears that 
these studies described are evaluating significant differences rather 
than statistical equivalence. As such, these pharmacological endpoint 
studies are not as rigorously designed from a statistical standpoint as 
classic bioequivalence plasma level studies, inasmuch as differences 
are being evaluated rather than equivalence. The comment suggested that 
pharmacological endpoint studies should also be evaluating statistical 
equivalence, rather than significant differences. In fact, a comparable 
equivalence testing is alluded to on page 22 regarding clinical 
endpoint studies, studies which would be expected to be less able to 
prove equivalence than pharmacologic endpoint studies.
    CVM agrees with the comment and has modified the guidance to read 
as follows:
    For parameters which can be measured over time, a time vs effect 
profile is generated, and equivalence is determined with the method 
of statistical analysis essentially the same as for the blood level 
bioequivalence study.
    For pharmacologic effects for which effect vs time curves can 
not be generated, then alternative procedures for statistical 
analysis should be discussed with CVM prior to conducting the study.
    32. Page 23, section VI. Human Food Safety Considerations. One 
comment asked if there is a need for determining a full depletion 
profile for the generic? The sponsor proposed that a single point 
tissue residue study completed out to the withdrawal time of the 
pioneer would be sufficient.
    The Center does not agree with the use of a single point tissue 
residue study at the withdrawal time of the pioneer as a general 
practice.
    A traditional tissue residue depletion study has always been 
required for generic products where bioequivalence is determined with a 
pharmacological or clinical endpoint study. The need for a traditional 
tissue residue depletion profile is expanded in the revised guidance to 
include blood level bioequivalence studies, because the Center has 
concluded that, with the exception of those examples listed in section 
VI. of the guidance, the tissue residue depletion of the generic 
product is not adequately addressed through bioequivalence studies.
    The use of the traditional tissue residue depletion study provides 
the Center with the data needed to compute a withdrawal period for the 
drug product in question, using our statistical tolerance limit model, 
whereby the 99th percentile is calculated with 95 percent confidence. 
Use of a single point tissue residue study ordinarily would not provide 
the data needed to use our current model, since the single-point study 
would not contain sufficient information regarding the variability of 
the residue depletion profile. Additionally, since the analytical 
methods approved for regulatory purposes can rarely measure the marker 
residue at the withdrawal time, a single point residue study at the 
pioneer withdrawal time would be limited by the efficiency of the 
regulatory analytical method at the drug concentrations typically seen 
at the pioneer withdrawal time. When the tissue residue values include 
negative or zero values (i.e., values below the limit of quantitation 
for the assay), the number of animals needed in the study will depend 
on the method variance and the number of zero values, and will vary 
from drug to drug. It is not possible to predict, a priori, the number 
of animals that will be needed to provide data of sufficient confidence 
for a single point tissue residue depletion study to obtain the 
confidence similar to that seen for the pioneer drug using our 
traditional residue depletion study design.
    The Center will consider the use of a single point tissue residue 
depletion study in those cases where the regulatory analytical method 
can be validated and demonstrated to measure reliably residues in the 
treated animals at the pioneer withdrawal time so that a 99th 
percentile statistical tolerance limit with 95 percent confidence can 
be calculated.
    A person may follow the guidance or may choose to follow alternate 
procedures or practices. If a person chooses to use alternate 
procedures or practices, that person may wish to discuss the matter 
further with the agency to prevent an expenditure of money and effort 
on activities that may later be determined to be unacceptable to FDA. 
Although this guidance document does not bind the agency or the public, 
and it does not create or confer any rights, privileges, or benefits 
for or on any person, it represents FDA's current thinking on 
bioequivalence testing for animal drugs. When a guidance document 
states a requirement imposed by statute or regulation, the requirement 
is law and its force and effect are not changed in any way by virtue of 
its inclusion in the guidance.
    Interested persons may, at any time, submit to the Dockets 
Management Branch (address above) written comments on the document. Two 
copies of any comments are to be submitted, except that individuals may 
submit one copy. Comments are to be identified with the docket number 
found in brackets in the heading of this document. The documents and 
received comments are available for public examination in the Dockets 
Management Branch between 9 a.m. and 4 p.m., Monday through Friday.

    Dated: May 17, 1996.
William K. Hubbard,
Associate Commissioner for Policy Coordination.
[FR Doc. 96-13106 Filed 5-23-96; 8:45 am]
BILLING CODE 4160-01-F