CVM GFI #267 Biomarkers and Surrogate Endpoints in Clinical Studies to Support Effectiveness of New Animal Drugs

Administrative guidance, not enacted legislation. This captured publisher edition preserves the listed document and attachments; later developments are not independently consolidated.

#267
Biomarkers and Surrogate Endpoints
in Clinical Studies to Support
Effectiveness of New Animal Drugs
Guidance for Industry
Submit comments on this guidance at any time. Submit electronic comments to
https://www.regulations.gov. Submit written comments to the Dockets Management Staff
(HFA-305), Food and Drug Administration, 5630 Fishers Lane, Rm. 1061, Rockville, MD
20852. All comments should be identified with the docket number FDA-2020-D-1402.
For further information regarding this document, contact AskCVM@fda.hhs.gov.
Additional copies of this guidance document may be requested from the Policy and Regulations
Staff (HFV-6), Center for Veterinary Medicine, Food and Drug Administration, 7500 Standish
Place, Rockville MD 20855, and may be viewed on the Internet at either
https://www.fda.gov/animal-veterinary or https://www.regulations.gov.
U.S. Department of Health and Human Services
Food and Drug Administration
Center for Veterinary Medicine (CVM)
October 2021

Contains Nonbinding Recommendations
2
Table of Contents
I. 	Introduction ....................................................................................................................... 3
II. 	Background ....................................................................................................................... 3
III. 	Scope................................................................................................................................... 4
IV. 	Biomarkers and Surrogate Endpoints in Veterinary and Human Medicine .............. 5
A. Regulatory Comparisons ............................................................................................ 5
B. Biomarkers as Drug Development Tools .................................................................. 6
C. Biomarkers as Surrogate Endpoints ......................................................................... 6
V. 	Explanation of Applicable Terms for Use within this Document ................................. 6
A. Biomarker .................................................................................................................... 6
B. Validation..................................................................................................................... 7
C. Endpoint and Outcomes ............................................................................................. 8
D. Surrogate Endpoints ................................................................................................... 9
VI. 	Analytical Considerations ................................................................................................ 9
VII. 	Biomarker Use in Effectiveness Studies........................................................................ 13
A. Context of Use ........................................................................................................... 13
B. Considerations for Different Stages of Drug Development................................... 16
C. Statistical Considerations ......................................................................................... 20
VIII. 	Labeling Considerations................................................................................................. 20
IX. 	Establishing a Biomarker as a Surrogate Endpoint for Effectiveness Studies ......... 21
X. 	Statistical Considerations for Studies to Establish a Biomarker as a Surrogate
Endpoint........................................................................................................................... 23
XI. 	Obtaining CVM Feedback on Use of Biomarkers and Surrogate Endpoints ........... 25
A. When to submit information regarding the use of biomarkers and surrogate
endpoints .................................................................................................................... 25
B. How to submit information regarding the use of biomarkers and surrogate
endpoints .................................................................................................................... 26
XII. 	References ........................................................................................................................ 27

Contains Nonbinding Recommendations
3
Biomarkers and Surrogate Endpoints in Clinical Studies to Support
Effectiveness for New Animal Drugs
Guidance for Industry
This guidance represents the current thinking of the Food and Drug Administration (FDA or
Agency) on this topic. It does not establish any rights for any person and is not binding on FDA
or the public. You can use an alternative approach if it satisfies the requirements of the
applicable statutes and regulations. To discuss an alternative approach, contact the FDA staff
responsible for this guidance as listed on the title page.
I. 	Introduction
FDA is issuing this Guidance for Industry (GFI), as required under section 305 of the Animal
Drug and Animal Generic Drug User Fee Amendments of 2018 (Pub. L. 115-234), to assist
sponsors in incorporating biomarkers and surrogate endpoints into proposed clinical
investigation protocols and applications for new animal drugs under the Federal Food, Drug, and
Cosmetic Act (FD&C Act). Section 305 of Pub. L. 115-234, among other things, directed FDA
to hold a public meeting for interested parties to discuss innovative animal drug investigation
designs and to issue guidance addressing the incorporation of the use of such elements of
investigations as complex adaptive and other novel investigation designs, data from foreign
countries, real-world evidence (including ongoing surveillance activities, observational studies,
and registry data), biomarkers, and surrogate endpoints into clinical investigation protocols and
applications to support the effectiveness of new animal drugs.
The contents of this document do not have the force and effect of law and are not meant to bind
the public in any way, unless specifically incorporated into a contract. This document is
intended only to provide clarity to the public regarding existing requirements under the law.
FDA guidance documents, including this guidance, should be viewed only as recommendations,
unless specific regulatory or statutory requirements are cited. The use of the word should in
Agency guidances means that something is suggested or recommended, but not required.
II. 	Background
In the Federal Register of July 9, 2019 (84 FR 32749), FDA’s Center for Veterinary Medicine
(CVM) published a notice of a public meeting entitled “Incorporating Alternative Approaches in
Clinical Investigations for New Animal Drugs” giving interested persons until August 17, 2019,
to comment on the topics discussed at the public meeting and the questions published in the
meeting notice (84 FR at 32750-32751).1 On August 13, 2019, we published a notice
1 https://www.fda.gov/animal-veterinary/workshops-conferences-meetings/public-meeting-incorporating-alternative-
approaches-clinical-investigations-new-animal-drugs

Contains Nonbinding Recommendations
4
announcing the extension of the comment period to September 16, 2019 (84 FR 40071). CVM
received numerous comments on the topics discussed at the public meeting and the questions
published in the meeting notice and those comments were considered as draft guidance was
developed.
This document describes recommendations for designing, conducting, and reporting the results
for investigations or studies including biomarkers and/or surrogate endpoints to demonstrate
substantial evidence of effectiveness or a reasonable expectation of effectiveness of drugs
intended for use in animals and to support the approval of a new animal drug application
(NADA) or an application for conditional approval of a new animal drug (CNADA). This
guidance also provides information about obtaining feedback from CVM with respect to
incorporating biomarkers and/or surrogate endpoints in investigations and study protocols for
new animal drugs. Other centers within FDA, including the Center for Drug Evaluation and
Research (CDER), the Center for Biologics Evaluation and Research (CBER), and the Center for
Devices and Radiological Health (CDRH), have released draft and final guidance documents on
the topics of biomarkers and surrogate endpoints. This guidance document provides CVM’s
recommendations specific to investigations for animal drugs.
CVM will consider all biomarkers and surrogate endpoints in submissions to investigational new
animal drug (INAD) files, NADAs, and CNADAs to demonstrate substantial evidence of
effectiveness or a reasonable expectation of effectiveness. This guidance document provides
CVM’s recommendations specific to investigations for animal drugs.
Some concepts and language in the recommendations for animal drugs are intended to be similar
or the same as those in other guidance documents issued by FDA on the same or similar topics.
Because these recommendations are specific to investigations for animal drugs, they have been
tailored to the unique aspects of and considerations for animal drug development.
III. 	Scope
The purpose of this guidance is to describe how CVM intends to evaluate biomarkers, including
surrogate endpoints, to determine whether they may be used to support substantial evidence of
effectiveness for an NADA or reasonable expectation of effectiveness for a CNADA. In
addition, this guidance describes the process by which sponsors may obtain feedback from CVM
on technical issues related to the use of biomarkers before the submission of an application. This
guidance does not address the use of biomarkers, including surrogate endpoints, to support
technical sections of an application other than Effectiveness or Reasonable Expectation of
Effectiveness.
CVM considers biomarker and surrogate endpoint evidence from individual drug sponsors in
addition to evidence from the broader scientific, veterinary, and human medical communities, as
applicable, as part of a development program for a specific new animal drug. CVM does not
have, and is not proposing, a formal qualification2 program to assess and qualify biomarkers,
2 Qualification is “a determination by the Secretary that a drug development tool and its proposed context of use can
be relied upon to have a specific interpretation and application in drug development and regulatory review under
[the FD&C Act].” FD&C Act § 507(e)(7).

Contains Nonbinding Recommendations
5
including surrogate endpoints, for use in effectiveness studies for the approval of new animal
drugs, independent from their use in a specific new animal drug development program. Animal
drug sponsors or other stakeholders considering qualification of a biomarker or surrogate
endpoint should refer to the resources available for human drug or device development and
discuss with CVM their applicability to new animal drug development before starting the
qualification process.3, 4, 5
CVM encourages sponsors to explore a variety of design options in the planning of clinical
effectiveness study designs and to proactively discuss their considerations with CVM. Some
recommendations in this guidance may be highly technical and we recommend you consult with
the appropriate CVM experts to help facilitate use of these principles.
IV. 	Biomarkers and Surrogate Endpoints in Veterinary Medicine
A. Regulatory Framework
An NADA must include, among other things, a “scientific rationale and purpose the new
animal drug is to serve: (a) Clinical purpose.”6 In this guidance, clinical purpose is
equated with “intended use,” which we have described as the disease or condition to be
treated, prevented, mitigated, or cured or the structure or function of the body to be
affected.7 Depending on the intended use, the new animal drug’s clinical purpose may
provide either clinical benefits or biological effects or both. A clinical benefit is defined
as a clinically meaningful effect of a drug on how an animal feels, functions, or survives.
A biological effect is defined as a meaningful effect on the structure or function of the
body or a clinically relevant pathophysiologic manifestation or characteristic of a disease
or condition. Sponsors must submit substantial evidence of effectiveness consisting of
one or more adequate and well-controlled studies to demonstrate that a new animal drug
is effective for each intended use and associated conditions of use.8 Generally, a sponsor
should use endpoints that provide direct evidence of the effectiveness of the intervention
with respect to the intended use. This may include the use of biomarkers alone or in
conjunction with clinical or biological outcomes as primary variables or, where
appropriate, the use of a biomarker as a surrogate endpoint. Surrogate endpoints should
predict either the intended clinical benefit or biologic effect.
3 https://www.fda.gov/drugs/drug-development-tool-qualification-programs/cder-biomarker-qualification-program
4 https://www.fda.gov/medical-devices/science-and-research-medical-devices/medical-device-development-tools-
mddt
5 See Draft GFI, “Biomarker Qualification: Evidentiary Framework,” (December 2018). When final, this guidance
will represent FDA’s current thinking on this topic.
6 21 CFR 514.1(b)(2)(ii)
7 See 62 FR 59830 at 59831 (November 5, 1997), which describes the characteristics of substantial evidence of
effectiveness.
8 21 CFR 514.4(b)(2)

Contains Nonbinding Recommendations
6
B. Biomarkers as Drug Development Tools
Sponsors may choose to use biomarkers as a part of early animal drug development or
within clinical effectiveness studies. A context of use (COU) is a statement that fully and
clearly describes the way the biomarker is to be used and the drug product development-
related purpose of the use.9 Depending on its intended COU, the use of previously
accepted biomarkers, including those used as surrogate endpoints, may require limited
justification. However, for the use of novel biomarkers, including the use of previously
accepted biomarkers in a different species, the individual sponsor may be responsible for
providing evidence supporting all aspects of the proposed biomarker and its proposed
COU.
C. Biomarkers as Surrogate Endpoints
Sponsors of new animal drugs may propose to use a biomarker as a surrogate endpoint to
predict a clinical benefit or biological effect. If a surrogate endpoint is known to predict
a clinical benefit or biological effect, it could be used to support substantial evidence of
effectiveness. Biomarkers that have not been fully established as surrogate endpoints
may be used to demonstrate a reasonable expectation of effectiveness for conditional
approval of certain new animal drugs.10 The animal drug’s effectiveness would then be
verified in studies using an established outcome to demonstrate substantial evidence of
effectiveness. For sponsors seeking to establish a biomarker as a validated surrogate
endpoint, see section IX. Establishing a Biomarker as a Surrogate Endpoint for
Effectiveness Studies below.
V. 	Explanation of Applicable Terms for Use within this Document
The terms related to biomarkers and study endpoints have been used and understood in different
ways in the various contexts of animal drug development. This section of the guidance seeks to
provide consistent, clear definitions of terms related to biomarkers and clinical study endpoints
for new animal drug approvals.
A. Biomarker
CVM considers a biomarker to be a defined characteristic that is measured as an indicator
of normal biological processes, pathogenic processes, or responses to an exposure or
9 https://www.ncbi.nlm.nih.gov/books/NBK338448/#IX-C
10 The Minor Use and Minor Species Animal Health Act (21 U.S.C. 360ccc) allows for the conditional approval of
new animal drugs intended for use in minor species (those other than horses, dogs, cats, cattle, pigs, turkeys, and
chickens) or for minor uses in major species (i.e., MUMS drugs). The Animal Drug and Animal Generic Drug User
Fee Amendments of 2018 (ADUFA IV) expanded the conditional approval pathway to include new animal drugs
that are intended to treat a serious or life-threatening disease or condition or address an unmet animal or human
health need and for which it is determined that a demonstration of effectiveness would require a complex or
particularly difficult study or studies. The need to develop a biomarker for an effectiveness endpoint, such as a
surrogate endpoint, would qualify for expanded conditional approval assuming all other criteria are met. See CVM
GFI #261, “Eligibility Criteria for Expanded Conditional Approval of New Animal Drugs,” (July 2021).

Contains Nonbinding Recommendations
7
intervention, including therapeutic interventions. More simply, biomarkers are measures
that can help characterize a baseline state, a disease process, or a response to a treatment.
Therefore, they can reflect physiological states, disease characteristics or processes, or
pharmacological responses in animals. A biomarker does not assess how an individual
feels, functions, or survives, (i.e., it does not assess the clinical benefit). In contrast to a
biomarker, an assessment of how an individual feels, functions, or survives is referred to
as a clinical outcome assessment.11 Clinical outcome assessments can be made through
reports by a clinician, a patient, or a non-clinician observer or through a performance-
based assessment. Therefore, biomarkers are not themselves assessments of a clinical
outcome.
Biomarkers are organized by type and category. The biomarker source or material for
measurement determines the biomarker type (e.g., molecular, histologic, radiographic,
physiologic characteristic, etc.). The biomarker’s use determines the biomarker category
(e.g., susceptibility/risk, diagnostic, monitoring, prognostic, predictive,
pharmacodynamic/response, or safety). For example, a susceptibility/risk biomarker is
used to assess an apparently healthy individual’s risk of developing a disease or medical
condition and may be based on a molecular, histologic, radiographic, or physiologic
characteristic. Changes in biomarkers after treatment reflect a biological response to the
drug product and may predict or identify safety problems related to a drug candidate (a
safety biomarker) or reveal a pharmacological activity expected to predict an eventual
desired outcome from treatment (a pharmacodynamic/response biomarker). These also
may be based on a molecular, histologic, radiographic, or physiologic characteristic.
A biomarker test is an assessment system used to detect or measure a biomarker and
comprises three essential components: (1) source or materials for measurement; (2) an
assay for obtaining the measurement; and (3) method and/or criteria for interpreting those
measurements. Examples of biomarker tests include an instrument or method for
measuring physiologic variables such as blood pressure, heart rate, or body temperature;
a test for measuring an enzyme in blood; or a medical imaging tool used to measure
tumor size.
B. Validation
Adequate and well-controlled studies must use a method of selecting animals that
provides adequate assurances that the animals are suitable for the purposes of the study
(21 CFR 514.117(b)(5)) and methods to assess animal response (endpoints/outcomes)
that are well-defined and reliable to establish effectiveness relative to the proposed
clinical purpose (intended use) of the new animal drug (21 CFR 514.117(b)(8)). Use of
fit-for-purpose biomarkers is one way to meet these characteristics of an adequate and
well-controlled study. Fit-for-purpose is a conclusion that the level of validation
associated with a biomarker is sufficient to support its COU.
11 https://www.fda.gov/about-fda/clinical-outcome-assessments-coa-frequently-asked-questions#COADefinition

Contains Nonbinding Recommendations
8
Validation is a process to establish that the performance of a test, tool, or instrument is
acceptable for its intended purpose. Biomarker test validation (analytical validation) is
separate from validation of the biomarker itself (clinical validation). Validation of
biomarker tests is important because reliable tests help ensure the quality of the
biomarker data obtained. Analytical validation is a process to establish that the
performance characteristics of a test, tool, or instrument are acceptable in terms of its
sensitivity, specificity, accuracy, precision, and other relevant performance characteristics
using a specified technical protocol (which may include specimen collection, handling
and storage procedures). This is validation of the test’s, tool’s, or instrument’s technical
performance but is not validation of the item’s usefulness. (See section VI. Analytical
Considerations for additional discussion on this topic.) In contrast, clinical validation is
a process to establish that the test, tool, or instrument acceptably identifies, measures, or
predicts the concept of interest. For biomarkers this means establishing that the
biomarker’s relationship with the outcome of interest is acceptable for the proposed
COU. Analytical validation and clinical validation are distinct processes. However, the
two processes are iterative and dependent on one another. A reliable test should be used
to measure the biomarker. Through this iterative process, experience with the biomarker
and the biomarker test could lead to improvements in the technical performance of the
test and the understanding of the biomarker’s biological and clinical significance.
C. Endpoint and Outcomes
Within this guidance document, CVM considers a study outcome to be a measurable
characteristic (e.g., clinical outcome assessment, biological outcome assessment, or
biomarker) that is influenced or affected by an individual’s baseline state or an
intervention to evaluate what happens to individuals in a clinical trial; whereas endpoints
are the analyzed parameter intended to reflect an outcome of interest. Endpoints are
defined based on the type of assessments made, the timing of those assessments, the
assessment tools used, and possibly other details, as applicable, such as how multiple
assessments within an individual are to be combined. For example, for an analgesic drug
intended to control post-operative pain, the study outcome may be the pain scale, whereas
the endpoint encompasses how the pain scale is assessed, such as the classification of that
score into the analyzed parameter of “success” or “failure.” The most reliable study
outcomes are those that directly measure the clinical benefit or biological effect.
CVM considers a clinical outcome to be an outcome that describes or reflects how an
individual (i.e., the animal) feels, functions or survives. Clinical outcomes generally
correspond to therapeutic indications for the diagnosis, cure, mitigation, treatment, or
prevention of disease. Examples of clinical outcomes are clinical outcome assessments,
such as rating scales for pain or clinical scoring systems for bovine respiratory disease.
In some cases, biomarkers may be combined with clinical outcome assessments to
establish effectiveness of a new animal drug.
In the context of this guidance document, CVM considers a biological outcome to be an
outcome that describes or reflects an effect on the structure or any function of the body of
a target animal; a consumer’s response to or evaluation of products, such as meat or milk,
derived from the target animal; or a molecular, histologic, radiographic, or physiologic

Contains Nonbinding Recommendations
9
characteristic (i.e., biomarker) of a recognized disease or condition. Biological outcomes
may correspond to either production or therapeutic indications, depending on the
intended biological effect of the drug. Examples of biological outcomes include the
measurement of rate of weight gain for a production drug, measurement of the carriage or
shedding status of a human pathogen, assessment of meat aroma and flavor, or the
measure of the infection level of a gastrointestinal nematode using worm counts.
D. Surrogate Endpoints
For new animal drugs, CVM considers a surrogate endpoint to be an endpoint that is used
in clinical studies as a substitute for a direct measure of a biological or clinical outcome
(Myers et al., 2017). In other words, an effect on the surrogate endpoint predicts a
clinical benefit or biological effect based on epidemiologic, therapeutic,
pathophysiologic, or other scientific evidence rather than measure the effect directly.
Surrogate endpoints are used when a clinical or biologic endpoint might take a very long
time to study, or in cases where the relationship between the change in the clinical or
biological outcome and the change in the surrogate endpoint is well understood.
Generally, clinical studies are needed to show that surrogate endpoints can be relied upon
to predict, or correlate with, a clinical benefit or biological effect. A surrogate endpoint
is considered validated after a sponsor demonstrates a clear mechanistic rationale and
submits strong evidence that an effect on the surrogate endpoint predicts a clinical
outcome or biological effect. When establishing a biomarker as a surrogate endpoint, it is
important to have a high level of confidence in the biomarker test’s analytical
performance when confirming the relationship between a biomarker and the clinical
outcome or biological effect of interest, and generally, studies intended to confirm this
relationship should be conducted using a validated test.
Validation is especially important when a surrogate endpoint biomarker is the basis for a
conclusion about substantial evidence of effectiveness of a new animal drug. A
biomarker that fails to represent disease processes or structural or functional changes and
the intended effectiveness of the new animal drug may lead to erroneous decisions (either
for approval or not) in new animal drug evaluation, even if the test for the biomarker is
analytically valid. If the effect of treatment on the biomarker does not adequately reflect
the effect on the traditionally accepted clinical or biological outcome, erroneous
conclusions might be drawn about the safety or effectiveness of the new animal drug.
A reasonably likely surrogate endpoint is an endpoint supported by strong mechanistic
and/or epidemiologic rationale such that an effect on the surrogate endpoint is expected
to be correlated with an outcome of primary interest, but without sufficient clinical data
to show that it is a validated surrogate endpoint.
VI. 	Analytical Considerations
Academic researchers, pharmaceutical companies, and diagnostic companies have published
white papers and reviews in an effort to establish “best practices” and industry standards for
analytical validation of biomarkers. FDA’s guidance for industry, “Bioanalytical Method

Contains Nonbinding Recommendations
10
Validation,”12 incorporates recommendations only pertaining to the validation of assays to
measure in vivo biomarker concentrations in biological matrices such as blood or urine. Because
biomarkers can be used for a wide variety of purposes during drug development, a fit-for-
purpose (FFP) approach should be used when determining the appropriate extent of method
validation. The guidance suggests that the approach used for drug assays should be the starting
point for validation of biomarker assays, although the FDA realizes that some characteristics
may not apply or that different considerations may need to be addressed.
Not all biomarkers are evaluated using assays. Some are measured by imaging, rulers, scales, or
other measuring devices. Therefore, the analytical validation studies and performance
characteristics vary greatly according to the biomarker type (molecular, histologic, radiographic,
and physiologic characteristic) and technology of the biomarker test.13 The specifics of the
validation methodology will depend on the measurement technology, the biomarker, and COU.
However, in general, the validation of the analytical procedure should determine if the
measurement method is able to discriminate the change in the biomarker needed to support
clinical validation, and a validation should demonstrate that measured changes, as applicable, are
acceptable for the COU.
Specific recommendations for assessing the performance characteristics of biomarker
measurements other than assays are beyond the scope of this guidance. However, the analytical
validation process of a biomarker measurement method is similar regardless of the biomarker
type, technology used for biomarker measurement, or the proposed COU and generally includes
the following:
• 	Description of the biomarker measurand(s) and measurement approach;
• 	Technical protocols to limit introduction of variability and improve reliability;
• 	Performance characteristics of the measurement method; and
• 	Analytical validation to ensure performance under COU conditions is suitable (accurate and
reliable).
Unlike the predefined acceptance criteria established for small and large molecule
pharmacokinetic assays, determining assay acceptance criteria for biomarker assays is likely the
most challenging exercise for a biomarker assay validation (Piccoli and Sauer, 2019). It should
be emphasized that the acceptance criteria for biomarker assays will depend heavily on the
intended use of the assay and should be based on physiological variability as well (Lee et al.,
2006). Analytical validation includes evidence to demonstrate that the test provides accurate and
precise measurements, including the determination of a cutoff, if necessary, for the interpretation
of the biomarker results. The cutoffs should be defined before the biomarker test is analytically
validated. Sufficient data should be provided to adequately describe the performance
12 See GFI, “Bioanalytical Method Validation” (May 2018)
13 See footnote 5 on page 5

Contains Nonbinding Recommendations
11
characteristics of the tool. Commercial availability of a test does not automatically imply that it
has sufficient analytical validation for its proposed COU.
As stated above, the FFP approach with validation criteria that are appropriate for the intended
use of the resulting data should be used. Whenever appropriate, recommendations and
acceptance criteria in FDA’s 2018 Bioanalytical Method Validation GFI should generally be
considered during method validation. Accordingly, analytical method validation for biomarker
assays should include the following analytical parameters.
Accuracy
The accuracy, sometimes termed “trueness,” of an analytical procedure expresses the closeness
of the determined value to the value that is accepted either as a conventional true value or an
accepted reference value. It is calculated as (the determined value/the true value) x 100%.
Accuracy should be established using at least three independent chromatographic assay runs or
six ligand binding assay (LBA) runs; four to five quality control (QC) levels per run (lower limit
of quantification (LLOQ), low (L), middle (M), high (H), upper limit of quantification (ULOQ));
and three chromatographic assay replicates or five LBA replicates per QC level.
Precision
The precision of an analytical procedure expresses the closeness of agreement (i.e., degree of
scatter) among a series of measurements obtained from multiple sampling of the same
homogenous sample under the prescribed conditions. Precision encompasses repeatability and
reproducibility and is expressed as the percent coefficient of variation of a series of
measurements.
• 	Repeatability is the closeness of agreement between results of successive measurements of
the same samples carried out in the same laboratory under the same operating conditions
within short intervals of time (Lee et al., 2006).
• 	Reproducibility is the closeness of agreement of results measured under significantly
changed conditions (e.g., inter-laboratory or alternate vendor of a critical reagent) (Lee et al.,
2006).
Precision should be established using at least three independent chromatographic assay runs or
six LBA runs; four to five quality control (QC) levels per run (LLOQ, L, M, H, ULOQ); and
three chromatographic assay replicates or five LBA replicates per QC level.
Sensitivity
Sensitivity is defined as the lowest analyte concentration in the matrix that can be measured with
acceptable accuracy and precision (i.e., LLOQ).
Specificity
Specificity is the ability of the method to unequivocally assess the analyte in the presence of
other components that are expected to be present (e.g., impurities, degradation products, matrix

Contains Nonbinding Recommendations
12
components, etc.). The method specificity should be assessed for interference by cross-reacting
molecules, concomitant medications, bio-transformed species, etc. Sponsors should make a
scientific judgment about the need to assess these (and any other) potential interferences during
method development.
Selectivity
Selectivity is the extent to which the method can determine a particular compound in the
analyzed matrices without interference from matrix components. Selectivity should be
established by analyzing blank samples of the appropriate biological matrix from at least six
individual sources in chromatographic assays, or at least 10 individual sources in LBAs. For
LBAs, it is important to investigate any interference originating from structurally or
physiologically similar analytes (i.e., exogenous interference) or matrix effects (i.e., endogenous
interference). Matrix effects evaluation involves comparing calibration curves in multiple
sources of the biological matrix against a calibration curve in the matrix for parallelism (serial
dilution of incurred samples) and nonspecific binding.14
Parallelism
Parallelism is the extent to which the dose-response relationship between two materials (i.e.,
calibrator versus unknown specimens) is constant for the examined range of concentrations
(Piccoli and Sauer, 2019). The absence of suitable blank matrices means that many biomarker
assays use calibrators prepared in a substitute matrix that differs from the test sample matrix (Lee
et al., 2006). If study samples are available, parallelism between the calibration standard curve
and serially diluted study samples should be assessed to detect possible matrix effect or differing
affinities for metabolites (EMA, 2011). A minimum of four serial dilutions of each sample
should be performed (Piccoli and Sauer, 2019). Parallelism between dilution curves, where
dilution of test samples in the range of the calibration curve does not result in substantially
different extrapolated analyte concentrations, validates the use of the substitute matrix for
calibrator preparation. Results of these experiments may also define suitable dilution ranges
should dilution be necessary to alleviate matrix effects (Lee et al., 2006).
Quantification (Calibration) Range
Quantification (Calibration) Range is the range of concentrations, including the ULOQ and the
LLOQ that can be reliably and reproducibly quantified with accuracy and precision with a
concentration-response relationship.
The choice of a curve fitting model for calibration curves should be tailored to each analytical
method because of the wide variety of assay formats and analytes in biomarker applications (Lee
et al., 2006). The simplest regression model for chromatographic methods and a four- or five-
parameter logistic model for LBAs are usually used, but other models may be acceptable with
14 See footnote 12 on p. 10

Contains Nonbinding Recommendations
13
justification.15 The appropriate calibration model should be confirmed with at least six
validation runs, with a blank and at least six, non-zero calibrator levels analyzed per each
validation run.
Stability
During method development, the sponsor should determine the chemical stability of an analyte in
a given matrix under specific conditions for given time intervals, including the effects of sample
collection, handling, and storage of the analyte. The sponsor should assess autosampler,
benchtop, processed or extracted sample, freeze-thaw, stock solution, and long-term stability of
the analyte. The sponsor should assess the stability in the same matrix as that intended for in-
study samples; however, when the matrix is rare, the sponsor may explore the use of suitable
surrogate matrices. Stability determinations should be performed using at least three replicates at
L and H QC concentrations.
Sample Integrity
Results from biomarker assays are valid only if sample integrity is maintained from sample
collection through analysis. Early, consistent application of predefined sample collection and
handling techniques is especially important when such manipulations might affect sample and/or
biomarker integrity. Provision of a detailed sample collection and storage protocol and adequate
training of clinical trial site personnel are especially important when extraordinary measures are
necessary to assure analyte integrity (Lee et al., 2006).
In-study Use of Validated Biomarker Assays
The sponsor should ensure that the assay continues to perform as per predefined specifications in
each study run (i.e., to ensure the assay remains in control) (Lee et al., 2006), by use of
appropriate QC samples (typically L, M, and H QCs, in duplicate at each level); using the same
curve fitting method, weighting, and goodness-of-fit test determined during the validation stage;
by performing incurred sample reanalysis; and reanalysis of samples based on reasons described
in a preexisting standard operating procedure.16 The specific recommendations and acceptance
criteria for in-study use of bioanalytical methods in GFI, “Bioanalytical Method Validation,”
could be used as a general guidance. However as in the validation stage, these recommendations
may be modified with justification, depending on the specific type of bioanalytical method.
VII. 	Biomarker Use in Effectiveness Studies
A. Context of Use
The information and level of evidence needed to support the use of a biomarker as part of
the development plan to establish the effectiveness of a new animal drug will depend on
its COU. For a new animal drug, the COU includes two components: (1) the biomarker
15 See footnote 12 on p. 10
16 See footnote 12 on p. 10

Contains Nonbinding Recommendations
14
category; and (2) the biomarker’s proposed use in drug development. The biomarker
may have multiple COUs in a drug development plan. If a sponsor is considering the use
of a biomarker in an effectiveness study, CVM encourages sponsors to discuss their
proposed COU with CVM early in the development process. Early communication with
CVM is especially important if the biomarker is to be an outcome used as the basis of a
decision regarding substantial evidence of effectiveness.
1. 	Biomarker category:
Biomarkers can be disease-related or treatment-related and should be classified
according to the following list:
• 	diagnostic biomarker: a biomarker used to detect or confirm presence of a disease
or condition of interest to identify individuals with a subtype of a disease;
• 	monitoring biomarker: a biomarker measured serially for assessing the status of a
disease or medical condition or for evidence of exposure to (or the effect of) a
medical product or an environmental agent;
• 	pharmacodynamic/response biomarker: a biomarker used to show that a
biological response has occurred in an individual exposed to a medical product or
an environmental agent;
• 	predictive biomarker: a biomarker used to identify individuals more likely than
similar individuals without the biomarker to experience a favorable or
unfavorable effect from exposure to a medical product or an environmental agent;
• 	prognostic biomarker: a biomarker used to identify the likelihood of a clinical
event or disease recurrence or progression in individuals having the disease or
medical condition of interest;
• 	safety biomarker: a biomarker measured before or after an exposure to a medical
product or an environmental agent to indicate the likelihood, presence, or extent
of toxicity as an adverse event; and
• 	susceptibility/risk biomarker: a biomarker that indicates the potential for
developing a disease or medical condition in an individual not currently having
clinically apparent disease or the medical condition.
2. 	Proposed use in drug development
The proposed use in drug development (pertinent to the effectiveness plan) should
include a consideration of some or all of the following points, as appropriate.
a) 	Purpose of use in drug development (e.g., a diagnostic biomarker to define
enrollment criteria in a clinical study; a prognostic biomarker to support
enrichment of a field study for a particular disease; a safety biomarker to evaluate
drug-induced liver injury within a field effectiveness study; or a

Contains Nonbinding Recommendations
15
pharmacodynamic/response biomarker as an outcome to demonstrate drug
effectiveness). Sponsors should discuss the biomarker’s purpose of use in drug
development in the context of the proposed clinical purpose for the new animal
drug (e.g., indication for use). If a biomarker is being used as a primary endpoint
in an effectiveness study, the biomarker’s proposed use should include whether
the biomarker will be used alone or together with other biomarkers or clinical
endpoints or if the biomarker is being proposed as a surrogate endpoint.
b) 	Proposed stage of drug development (e.g., exploratory/discovery phase, dosage
characterization, reasonable expectation of effectiveness, and/or substantial
evidence of effectiveness).
c) 	Clinical study population or model system (e.g., healthy animals; animals
undergoing surgery or diagnosed with a particular disease; cultured white blood
cells; etc.).
d) 	Mechanism of action (MOA) of the new animal drug for which the biomarker is
intended to have value, provided that the MOA is relevant to the biomarker’s
biology, the biomarker’s proposed COU, and the intended effect of the new
animal drug (e.g., both the MOA and the biomarker are within the same biologic
pathway or process) (Fleming and Powers, 2012). For some biomarker uses, such
as a diagnostic biomarker used for enrollment or a prognostic biomarker used to
support enrichment of a field study for a particular disease, information about the
MOA of the new animal drug may not be a necessary component of the COU for
the biomarker. However, in situations where biomarkers are used as primary
effectiveness endpoints, including their use as surrogate endpoints or as a
predictive biomarker used to support enrichment of a field study for improved
drug effectiveness, information about the MOA of the new animal drug may
impact the evaluation of the biological or scientific plausibility of the biomarker’s
proposed COU.
Sponsors should consider, and in some cases submit to CVM for comment, an
assessment of the benefits and risks17 of the use of the biomarker for the proposed
COU, as well as any potential risk mitigation strategies. The overall balance of
benefits, risks, and risk mitigation efforts are important to determine the level of
evidence needed to support the use of the biomarker.
The potential benefits of a biomarker for use in drug development depend on the
biomarker’s proposed COU. Biomarker use could benefit study animals participating
in clinical trials (e.g., earlier identification of toxicity with a safety biomarker) or
general drug development and regulatory decision making (e.g., a prognostic or
17 The terms “benefit,” “risk,” and “risk mitigation” that are used in the context of biomarker evaluation for new
animal drugs have specific meanings that are relevant to biomarker development and evaluation, and these meanings
are separate and distinct from how these terms are used in the context of evaluating the safety and effectiveness of
medical products.

Contains Nonbinding Recommendations
16
predictive biomarker used to enrich a patient population might reduce the sample size
needed to achieve statistical significance).
The potential risks of using a biomarker in the effectiveness assessment of a new
animal drug should address the consequences of incorrect decision making or harm to
study animals if the biomarker does not perform as expected and should consider
factors that might mitigate harm. For example, if a pharmacodynamic/response
biomarker fails to accurately establish the effectiveness of a new animal drug, study
animals may be unnecessarily used in a clinical study and an effective drug may not
actually be approved. Alternatively, if a pharmacodynamic/response biomarker
without sufficient validation is used in a clinical study and appears to demonstrate
effectiveness, where effectiveness does not exist, there is the risk of approving an
ineffective new animal drug. These risks (e.g., inaccurately establishing
effectiveness) could be mitigated, in part, by using the proposed biomarker along
with clinical outcomes or other biological outcomes as part of a multi-component
endpoint, rather than as a stand-alone assessment for the establishment of
effectiveness. In another example, a prognostic biomarker intended for clinical study
enrichment might fail to identify study animals with more rapid disease progression.
In this case, the mitigation strategies might include incorporating an interim analysis
for sample size re-estimation.
Evidence to support biomarker use consists of data to support analytical validation
and clinical validation. These data may be derived from various sources, including
pilot studies, clinical trials, foreign studies, veterinary records, and other sources,
discussed further below. The level of evidence needed to use a biomarker in the
effectiveness evaluation for a new animal drug depends on the selection of the
appropriate biomarker for the proposed COU, the quality of the biomarker
measurement, and if necessary, the correlation of the biomarker with the outcome of
interest.
B. Considerations for Different Stages of Drug Development
Exploratory/discovery phase
During early drug development, sponsors may use biomarkers to identify reasons for
differences between responders and non-responders to a drug (e.g., using certain genetic
markers), which is important to inform later study designs, such as enrichment
strategies,18 and potentially the development of companion diagnostic tests; development
of prognostic information to understand the natural history of a disease; or to identify
potential disease targets for therapy.
Dosage characterization
18 See CVM GFI #268, “Adaptive and Other Innovative Designs for Effectiveness Studies of New Animal Drugs,”
(October 2021)

Contains Nonbinding Recommendations
17
For the purposes of dosage characterization, sponsors should submit a justification of the
dosage (dose or dose range, dosing frequency, and the dosing duration) and a
characterization of the critical aspects of the dose response relationship related to each
intended use and associated conditions of use. Dosage characterization information may
be derived from dose titration studies, pilot studies, foreign studies,19 or scientific
literature. A biomarker, such as a pharmacodynamic marker, may also be used to select
the dosage to be used in studies to demonstrate reasonable expectation of effectiveness or
substantial evidence of effectiveness. Although CVM does not expect the sponsor to
submit a discussion of the COU, or an assessment of the benefits and risks of the use of a
biomarker for the purposes of dosage characterization, sponsors should consider the
quality of the analytical and clinical validation, the reliability of the biomarker(s) used for
dosage characterization, and the impact of the biomarker use on the drug development
plan and on the confidence in the dosage selection for use in studies to demonstrate
reasonable expectation or substantial evidence of effectiveness.
Reasonable expectations of effectiveness
CVM would not expect biomarkers used in studies to demonstrate reasonable expectation
of effectiveness (for conditional approval, as described in section IV. Biomarkers and
Surrogate Endpoints in Veterinary Medicine of this guidance) to have the same level of
analytical and/or clinical validation as biomarkers used as primary endpoints in studies
used to demonstrate substantial evidence of effectiveness. However, insufficient
analytical and/or clinical validation could lead to a failure to accurately predict
effectiveness in studies used to generate substantial evidence of effectiveness.
A need to develop and “qualify” effectiveness endpoints such as biomarkers to conduct
the clinical study to demonstrate substantial evidence of effectiveness may qualify a
project for the expanded conditional approval pathway if all other requirements for
conditional approval are met.20
Substantial evidence of effectiveness
Biomarkers may be used in a variety of ways in studies designed to demonstrate
substantial evidence of effectiveness including providing information to inform the
design of the study or serving as a primary endpoint, including their use as a validated
surrogate endpoint.
1. 	Use of biomarkers in the design of clinical effectiveness studies
Biomarkers may be used to determine whether an animal meets enrollment criteria
(e.g., to confirm a diagnosis of disease in enrolled animals), to refine the population
19 See CVM GFI #265, “Use of Data from Foreign Investigational Studies to Support Effectiveness of New Animal
Drugs,” (October 2021)
20 See CVM GFI #261, “Eligibility Criteria for Expanded Conditional Approval of New Animal Drugs,” (July
2021).

Contains Nonbinding Recommendations
18
to enroll (e.g., to enrich for specific study population), or to establish appropriate
randomization factors (e.g., appropriate allocation of animals at various disease
stages across treatment groups). For example, a threshold fecal egg count value may
be used to include or exclude animals in an anthelmintic dose confirmation study;
radiography may be used to confirm a diagnosis of osteoarthritis in dogs for an
analgesic study; and special stains used at the time of histopathology may be used to
enrich for a specific study population (e.g., proliferative index of mast cell tumors).21
2. 	Biomarkers as primary endpoints in clinical effectiveness studies
Any method used to assess an animal’s response to treatment in a clinical study,
including biomarkers, must be well-defined and reliable (21 CFR 514.117(b)(8)) and
be used in such a way to allow for a conclusion that the measured parameters and
responses reliably reflect the effectiveness of the new animal drug.
Depending on the clinical purpose of the drug, as reflected in the indication (the
intended use) and associated conditions of use, effects on a biomarker may be used as
a primary endpoint in conjunction with clinical outcomes to demonstrate biological
effects or clinical benefit, as the sole primary endpoint to demonstrate certain
biological effects, or as a surrogate endpoint to predict biological effects or clinical
benefit. For biomarkers used as primary outcome variables, including as surrogate
endpoints, appropriate statistical analysis methods should be chosen according to the
distribution of the biomarker following the general principles of statistics.
a. 	Use of biomarkers in conjunction with clinical or biological outcomes
Biomarkers that have not been validated as a biological outcome for a biological
effect or as a surrogate endpoint to substitute for a clinical outcome to predict a
clinical benefit should not be used as the only outcome for the demonstration of
substantial evidence of effectiveness. However, it may be appropriate to use the
biomarker(s) in conjunction with clinical outcomes or other biological outcomes,
depending on the COU. An example is the use of serum total T4 levels in
conjunction with improvement in clinical signs to evaluate the effectiveness of
antithyroid drugs for the treatment of hyperthyroidism in cats; and the use of the
Na+ and K+ levels or the Na+ /K+ ratio along with resolved or reduced clinical
signs to evaluate the effectiveness of mineralocorticoid drugs for use as
replacement therapy for mineralocorticoid deficiency in dogs with primary
hypoadrenocorticism.
b. 	Use of biomarkers as the sole outcome to demonstrate a biological effect
In some cases, a biomarker that is not validated as a surrogate endpoint may be
used as a biological outcome, and effects on the biomarker serve as the primary
21 See section V.A. Enrichment Strategies of CVM GFI #268, “Adaptive and Other Innovative Designs for
Effectiveness Studies of New Animal Drugs,” (October 2021)

Contains Nonbinding Recommendations
19
endpoint for a biological effect indication. In these situations, the purpose of the
new animal drug (i.e., the biological effect) should have appropriate utility to the
end user.
An example is the measurement of ammonia gas emissions (biomarker/biological
outcome) from cattle in chambers to demonstrate the effectiveness of a new
animal drug to reduce ammonia gas emissions (the biological endpoint) in cattle
in feedlots. Another example is the use of nematode worm counts to demonstrate
the effectiveness of a drug to treat and control gastrointestinal parasite infections.
In this situation, worm counts are evaluated to demonstrate a biological effect that
has utility to the veterinarian and end user (treatment and control of the parasite
species), rather than a clinical outcome (such as clinical signs of diarrhea, anemia,
or weight loss) to demonstrate a clinical benefit.
c. 	Use of biomarkers as surrogate endpoints
For some clinical studies, the development and use of a biomarker as a substitute
(surrogate) for a clinical outcome or certain types of biological outcomes may be
desirable. If a surrogate endpoint biomarker can be measured earlier or more
easily than the clinical outcome (such as mortality or a pain score, respectively), it
may help avoid studies of excessively long duration or with invasive or difficult-
to-obtain endpoints. For example, in veterinary oncology, a consensus framework
using the human response evaluation criteria in solid tumors (RECIST v1.1)
outlined by the Veterinary Cooperative Oncology Group provides guidelines for
the use of response criteria that may be used as surrogate endpoints for survival in
the evaluation of treatment effectiveness of solid tumors in dogs (Nguyen et al.,
2015). In another example, CVM accepted sperm count as a surrogate endpoint
for male sterility as a substitute for an evaluation of the dog’s ability to reproduce
clinically, for the approval of a drug indicated for chemical sterilization in dogs.22
Biomarkers proposed as surrogate endpoints in studies to demonstrate substantial
evidence of effectiveness should be validated, such that they accurately represent
and predict clinical effectiveness, are analytically valid, and fit the specific
intended context of the proposed indication (i.e., its intended COU). The
evidence used to validate a biomarker as a surrogate endpoint for substantial
evidence of effectiveness should be generated separately from the data used to
determine effectiveness for a specific drug. Use of a surrogate endpoint does not
change the requirement for substantial evidence of effectiveness but offers an
alternate method of measuring effectiveness.
3. 	Biomarkers as secondary endpoints
Biomarkers used as secondary endpoints (i.e., ancillary information that may be
supportive) may not be fully validated but they should always be shown to be fit-for-
22 NADA 141-217, available on Animal Drugs @FDA

Contains Nonbinding Recommendations
20
purpose. Biomarkers that are used as part of a multi-component primary endpoint are
not considered secondary endpoints.
C. Statistical Considerations
For surrogate endpoints that need to be established through studies, we recommend that
you refer to section IX. Establishing a Biomarker as a Surrogate Endpoint for
Effectiveness Studies of this guidance for statistical considerations and consult with FDA
statisticians regarding the detailed analysis plan you will be undertaking in your study.
When biomarkers are used as part of a multi-component primary endpoint to demonstrate
substantial evidence of effectiveness, the basis of study conclusion and analysis should be
prospectively specified to avoid a problem with multiplicity.23
VIII. 	Labeling Considerations
Determining the indication may be straightforward for many intended uses (e.g., the disease,
condition, manifestation, or clinical signs of the disease or condition being treated, prevented,
mitigated, cured, or diagnosed; the structural or functional changes; or alterations in clinically
relevant pathophysiologic manifestations of a disease or condition). In such circumstances, the
indication should convey the clinical benefit or biologic effect. For prescription new animal
drugs, the descriptions of specific outcomes and endpoints should be summarized in the
effectiveness section of labeling and should not be included in the indication.
When biomarkers and surrogate endpoints are used as primary variables to demonstrate the
effectiveness of a drug for an intended use, they are typically not described in the labeling except
for within the description of the effectiveness studies. However, there are three circumstances
when identifying a biomarker or surrogate endpoint in the indication may be appropriate.
First, if a drug only affects certain manifestations of a disease or condition or biological outcome
and this effect was directly measured using a biomarker, the indication should be worded to
reflect the intended outcomes with respect to that biomarker.
Second, it may be appropriate to include additional context about the approval in the indication
by identifying the clinical benefit(s) or biological effect(s) that are expected (based on the effect
demonstrated on the surrogate endpoint) but not yet established. Such circumstances should be
determined on a case-by-case basis and are most common with CNADAs because they are
approved based on a reasonable expectation of effectiveness rather than substantial evidence of
effectiveness.
Third, in some cases, a broad indication for a disease or biological effect may not be appropriate
and an indication identifying an outcome or endpoint, including a biomarker or surrogate
endpoint, may be considered. Such cases include when the drug’s effect on the overall disease or
biological outcome is not well understood; when different drugs have different effects on various
23 See Draft GFI, “Multiple Endpoints in Clinical Trials,” (January 2017). When final, this guidance will represent
FDA’s current thinking on this topic.

Contains Nonbinding Recommendations
21
manifestations of a disease or aspects of a biologic outcome; when studies evaluated only one or
some of the manifestations of the disease or aspects of a biologic outcome; or when the
endpoints are different from typical effectiveness measures.
IX. 	Establishing a Biomarker as a Surrogate Endpoint for Effectiveness Studies
For a biomarker development effort to be successful, the biomarker should be clearly identified
and characterized, including its source material or matrix and its method of measurement. The
biomarker should be clearly identified based on the specific analyte (e.g., fibrinogen), anatomic
feature (e.g., joint angle), or physiological characteristic (e.g., blood pressure) being measured.
For composite biomarkers, it is important to list the individual biomarker components and how
these components are interrelated (e.g., a description of an algorithm or scoring system). If
individual components have differential weighting, the description should include the biologic
rationale to support this decision. Because biomarkers are measured entities, it is important to
describe the biomarker source or material for measurement, which determines the biomarker
type (e.g., molecular, histologic, radiographic, or physiologic characteristic). For example, a
molecular biomarker obtained from a biofluid should state the sample matrix (e.g., plasma, urine,
etc.), and a radiographic biomarker should include the organ or tissue imaged (e.g., kidney).
The items that should be considered when determining the type and level of evidence sufficient
to support the use of a biomarker as a surrogate endpoint consists of several components. The
items include: (1) describing the drug development need; (2) defining the COU; (3) considering
potential benefits if the biomarker is accepted for use; and (4) considering potential risks
associated with the proposed use of the biomarker in a drug development program.
Ultimately, whether there is sufficient evidence to support acceptance of a biomarker as a
surrogate endpoint depends on the selection of the appropriate biomarker for the proposed COU,
the quality of the biomarker measurement, and the correlation of the biomarker with the outcome
of interest. Evidence to support biomarker acceptance consists of data to support clinical
validation and analytical validation. As defined above, clinical validation is the process to
establish that the test, tool, or instrument acceptably identifies, measures, or predicts the concept
of interest. Clinical validation establishes that a biomarker’s relationship with the outcome of
interest is acceptable for the proposed COU.
The sponsor should describe what is known about the biomarker’s role in the causal or outcome
pathway of interest, as well as describe knowledge gaps about the pathophysiology and
molecular underpinnings of the disease, condition, or outcome of interest. Describing the
biomarker’s position in the physiologic or pathologic pathway, if applicable, helps to support the
biological plausibility of the biomarker’s role in the proposed COU. The sponsor should provide
data supporting the relationship between the biomarker and the outcome of interest. This
relationship should be supported by statistical analyses (see section X. Statistical Considerations
below) and should come from multiple independent data sources, if possible. Together this
information can establish the clinical validity of a biomarker for a specified COU.
When assessing whether the association between a biomarker and an outcome of interest is
acceptable for the proposed COU, a key consideration is how to define the outcome of interest.
In some settings, there might not be a current standard outcome, or a standard outcome with

Contains Nonbinding Recommendations
22
known limitations is used for comparative purposes. For example, changes in serum creatinine
are widely used in biomarker development as the current standard for predicting changes in
kidney function. However, changes in serum creatinine levels are neither highly sensitive nor
highly specific for changes in kidney function. In a setting in which the current standard
outcome has significant limitations or a current standard outcome does not exist, it is important
to consider the totality of all available data that may provide sufficient information to establish
that the biomarker can be acceptably relied upon for the proposed COU.
Sponsors should provide evidence of the prognostic and/or predictive value of the biomarker as a
surrogate endpoint. Sponsors also should provide evidence and sufficient justification that the
biomarker test is analytically fit-for-purpose within the specific context of the proposed
indication. This evidence should be collected separately from the data that will be used to
support substantial evidence of effectiveness of a new animal drug. CVM encourages sponsors
to establish biomarkers for use as surrogate endpoints as early in the development phase as
possible, ideally prior to seeking concurrence on an effectiveness study protocol in which the
biomarker is used as a surrogate endpoint.
Sponsors should also provide evidence to support analytical validation of the biomarker for use
as a surrogate endpoint (see section VI. Analytical Validation for Biomarkers).
Sponsors should consider and address, as appropriate, the following regarding the use of
surrogate endpoints.
1. 	What is the outcome the surrogate endpoint is proposed to predict?
2. 	What is the rationale for using a surrogate endpoint rather than a clinical outcome or
traditionally measured biological outcome (e.g., feasibility, study duration, sample size,
etc.)?
3. 	What evidence exists to support the relationship between the surrogate endpoint and the
outcome of interest (e.g., epidemiologic studies, randomized controlled trials, real world
evidence generated using data from veterinary medical records, disease registries, and/or data
from mobile or remote sensing devices24; data generated from therapeutic products from the
same class, etc.)?
4. 	If the surrogate endpoint is proposed based on prior publications and general scientific
community acceptance, discuss the current body of evidence and the use of the surrogate
endpoint in clinical studies.
5. 	What is the relationship of the surrogate endpoint with the causal pathway for the disease or
the predicted biological effect?
6. 	How much do changes in the surrogate endpoint reflect changes in the outcome of interest or
the probability of the outcome occurring?
24 See CVM GFI #266, “Use of Real World Data and Real World Evidence to Support Effectiveness of New Animal
Drugs,” (October 2021)

Contains Nonbinding Recommendations
23
7. 	What is the extent and timing of change in the surrogate endpoint that would predict the
outcome of interest?
8. 	Is the change in the surrogate endpoint stable or does it only occur for a short time?
9. 	If a minimum threshold for change has been selected (both size and duration), how was this
value determined (include studies conducted and data generated)? What information is
available about the sensitivity and specificity of any measurement tools used?
10. How does the surrogate endpoint predict the outcome across different subgroups of the
targeted population (e.g., animal class, disease severity, co-existing conditions, concomitant
medications typical of the population, etc.)?
11. Is there data showing that the surrogate endpoint predicts the response similarly across
relevant subgroups of the targeted population?
12. What is the evidence that a drug-induced change in the surrogate endpoint will be predictive
of a change in the outcome (e.g., the surrogate endpoint is a correlate vs. the surrogate
endpoint has actual predictive value)?
X. 	Statistical Considerations for Studies to Establish a Biomarker as a Surrogate
Endpoint
The goal of statistical analyses to establish a biomarker as a surrogate endpoint is to evaluate the
degree and certainty of association between a biomarker and the outcome of interest. This
subsection describes the general considerations in the study design, sources of data, statistical
analysis methods in assessing the biomarker, and the criteria to establish the biomarker as a
surrogate endpoint.
Statistical evidence for establishing a biomarker as a surrogate endpoint is usually based on trial-
level analysis. Patient-level analysis can be used as secondary supporting evidence. Some of the
statistical approaches for assessment of biomarker surrogacy include the following:
• 	The strongest level of evidence to support the association of a biomarker with an outcome of
interest comes from prospective controlled studies that are specifically designed to assess the
association. In designing a study to demonstrate the association between a biomarker and an
outcome, many of the general statistical principles for clinical trials, such as randomization,
masking, sample size, and multiplicity issues, are relevant and useful. However, special
considerations should be taken as follows:
o 	The purpose of such studies is to establish the association between a biomarker and an
outcome, not to test for significant differences. Therefore, the correlation and its
confidence interval may be more useful, and over-reliance on p-values should be avoided,
unless the study is designed to test the null hypothesis of no relationship of the biomarker
to the outcome of interest.
o 	When multiple candidate biomarkers are evaluated in one study, the potential for false
positive results is increased; therefore, the analysis plan should control for multiplicity.

Contains Nonbinding Recommendations
24
• 	Systematic review may be used to evaluate the association of a biomarker with an outcome.
Systematic review is a process where data from multiple studies, including those published in
the scientific literature, are identified and evaluated, with a prospective study design and
analysis plan.
• 	Meta-Analysis based on published data is another approach to evaluate the association of a
biomarker with an outcome. The general statistical principles for performing meta-analysis
should be followed.
• 	Other innovative statistical approaches, such as a Bayesian method including prior
information and hierarchical models, may be considered to evaluate the association of a
biomarker with an outcome. As with any innovative statistical proposal, CVM recommends
sponsors discuss with the appropriate CVM review division the approaches to be undertaken
in the study.
Data used to assess the relationship between a biomarker and an outcome of interest may come
from a variety of sources. Some examples are as follows:
• 	Randomized, controlled clinical trials;
• 	Open-label/historical-control clinical trials; and
• 	Observational studies, such as longitudinal cohort studies, case-control studies, cross-
sectional data, and case reports.
The sponsor should develop a detailed plan on how the data will be collected and selected for the
intended use. When published data are used, certain limitations and gaps, such as publication
bias and data quality assurance, should be considered and specified a priori.
An appropriate statistical analysis method for quantitative evaluation of the relationship between
the biomarker and the biological or clinical outcome should be pre-specified in the study
protocol. Typically, the analysis plan should include definitions of the biomarker variable and
the biological or clinical outcome, as well as the statistical analysis model. The analysis method
should be consistent with the study design and the distribution properties of the variables of
interest. For example, for continuous measures, results can be evaluated using regression
models. For binary variables, results can be evaluated using sensitivity and specificity. When
continuous measurements are categorized, the level of cutoff values should be clinically
meaningful.
The success criteria to establish the biomarker as a surrogate endpoint should be pre-specified.
There are no standard quantitative criteria for determining whether the relationship between the
biomarker and the biological or clinical outcome is sufficiently strong to support surrogacy. The
evidence of strong association may vary, depending on the characteristics of the biomarker and
the biological or clinical outcome. The sponsor should discuss with CVM the criteria before the
study is conducted.

Contains Nonbinding Recommendations
25
XI. 	Obtaining CVM Feedback on Use of Biomarkers and Surrogate Endpoints
There are various approaches that sponsors may take to open a discussion with CVM on the use
of biomarkers and surrogate endpoints as part of their development program to demonstrate
effectiveness or a reasonable expectation of effectiveness. The sponsor’s decision regarding
which approach to select may be affected by where the project is in the development process.
Communication about biomarkers and surrogate endpoints may occur at any point in the
development process.
The Office of New Animal Drug Evaluation (ONADE) project managers (PMs) serve as a
central point of contact for drug sponsors and can provide information about the new animal
drug review process and ONADE’s regulatory procedures. If you have questions about the
approval process and do not have an ONADE PM assigned to your company, you can contact the
PMs through the CVM mailbox AskCVM@fda.hhs.gov.
A. When to submit information regarding the use of biomarkers and surrogate
endpoints
There are a variety of points in the development process and a variety of submission
types that can be used to obtain feedback. CVM encourages sponsors interested in using
biomarkers and surrogate endpoints as part of their development program for a new
animal drug to inform CVM as early in the product development process as possible.
Sponsors planning to incorporate biomarkers and surrogate endpoints to demonstrate
effectiveness or reasonable expectation of effectiveness are encouraged to inform CVM
of their intent either as part of their initial request to open a General Correspondence
(GC) file or an INAD file (A-0000), or as part of their initial presubmission conference
with CVM to discuss the drug product development plan (Z-submission product
development meeting). If one or more studies incorporating biomarkers and surrogate
endpoints are already complete, CVM recommends sponsors submit the information
described in section XI.B. How to submit information regarding the use of biomarkers
and surrogate endpoints prior to the initial presubmission conference. While CVM
cannot make a determination if existing data satisfies technical section requirements
outside of a data submission, if the sponsor submits sufficient information about existing
data early, we can provide feedback to help inform the development plan.25 Sponsors are
also encouraged to contact their assigned PM for assistance in determining the most
appropriate method for obtaining feedback from CVM.
25 See CVM Program Policy and Procedures (P&P) Manual 1243.2200 Submission and Review of Early
Information (EI) Prior to Presubmission Conferences and Protocol Review (June 2020) and CVM P&P Manual
1243.3050 Determining Technical Section Requirements for New Animal Drug Product Approval (May 2019)

Contains Nonbinding Recommendations
26
B. How to submit information regarding the use of biomarkers and surrogate
endpoints
There are several ways that sponsors may submit detailed information about plans for
incorporating biomarkers and surrogate endpoints into their development program to
demonstrate substantial evidence of effectiveness or reasonable expectation of
effectiveness. The regulatory pathway selected (CNADA versus NADA), the stage of
development, the information available, and the feedback being sought from CVM,
among other factors, may influence the submission type selected.
Sponsors may seek general guidance on the use of biomarkers and surrogate endpoints to
a GC file prior to opening an INAD file. Sponsors may submit information to support
use of data from biomarkers and surrogate endpoints as part of their initial request to
open an INAD file; as part of a meeting request for a presubmission conference (Z-
submission) to discuss the Effectiveness technical section requirements; or as part of an
information submission (H-submission) or meeting request (Z-submission) to discuss
study protocol design.
Sponsors considering incorporating biomarkers and surrogate endpoints into future
studies to demonstrate effectiveness or reasonable expectation of effectiveness should,
prior to conducting a study, submit a study protocol for review (E-submission).
Obtaining CVM input regarding study design will make reaching protocol concurrence
more efficient.
Sponsors may also open a Veterinary Master File (VMF) to hold detailed information
regarding a specific study design, including those regarding pre-investigational
discussions about the use of biomarkers and surrogate endpoints, or if the information
will be used in the development of multiple applications.26 The VMF is confidential and
is typically used when a holder wishes the material in the VMF to remain proprietary,
although the material may be referenced by multiple third-party products or files (INAD,
NADA, or CNADA). Alternatively, if multiple sponsors are cooperating on product
development, sponsors may establish a Public Master File (PMF) to allow all cooperators
to reference the information. As suggested by the name, the information in a PMF is
publicly available.
Regardless of how information is submitted to CVM, sponsors should submit an
organized and focused information package. This will allow CVM the best opportunity
to provide appropriate recommendations in response. Although full information may not
be available in the early stages of the development process, the amount of information
provided and the level of detail of the information provided should be commensurate
with the submission type. The information should address some or all of the following
elements, as appropriate for the submission type:
26 See CVM P&P Manual 1243.2400 Veterinary Master Files with Manufacturing Information (August 2019)

Contains Nonbinding Recommendations
27
1. 	The proposed COU, including the category of biomarker and its proposed use in
biomarker development;
2. 	An assessment of the benefits and risks27 of the use of the biomarker for the proposed
COU;
3. 	A description of the sponsor’s plan for analytical and clinical validation, as
appropriate;
4. 	A discussion of known prior uses of the biomarker (e.g., in clinical practice settings,
other countries, etc.); and
5. 	A statement regarding whether other FDA Centers have qualified the biomarker for
the same or similar purposes in the target animal or other species.
Additional questions and considerations specific to surrogate endpoints are presented in
section IX. Establishing a Biomarker as a Surrogate Endpoint for Effectiveness Studies.
XII. 	References
EMA, 2011. European Medicines Agency Guideline on Bioanalytical Method Validation
available at: https://www.ema.europa.eu/en/documents/scientific-guideline/guideline-
bioanalytical-method-validation_en.pdf
Fleming, TR and JH Powers. 2012. Biomarkers and surrogate endpoints in clinical trials. Stat
Med 31(25): 2973-2984.
Lee, JW, V Devanarayan, YC Barrett, R Weiner, J Allinson, S Fountain, S Keller, I Weinryb, M
Green, L Duan, JA Rogers, R Millham, PJ O'Brien, J Sailstad, M Khan, C Ray, and JA Wagner.
2006. Fit-for-purpose method development and validation for successful biomarker
measurement. Pharm Res no. 23 (2):312-28. doi: 10.1007/s11095-005-9045-3.
Myers, MJ, ER Smith, and PG Turfle. 2017. Biomarkers in Veterinary Medicine. Annu Rev
Anim Biosci no. 5:65-87. doi: 10.1146/annurev-animal-021815-111431.
Nguyen, SM, DH Thamm, DM Vail, and CA London. 2015. Response evaluation criteria for
solid tumours in dogs (v1.0): a Veterinary Cooperative Oncology Group (VCOG) consensus
document. Vet Comp Oncol 13(3): 176-183.
Piccoli, SP, and JM Sauer. 2019. Points to Consider Document: Scientific and Regulatory
Considerations for the Analytical Validation of Assays Used in the Qualification of Biomarkers
in Biological Matrices. edited by Biomarker Assay Collaborative Evidentiary Considerations
Writing Group.
27 See footnote 5 on page 5