AMGEN/media/Supporting Documents/T… · Amgen Inc. is pleased to respond to the FDA's draft guidance documents on the ... 2010 public hearing. Amgen looks forward to ... FDA Biosimilar

Submitted on the following website: http://www.regulations.gov

16 April 2012

Division of Dockets Management (HFA-305) Food and Drug Administrat ion 5630 Fishers Lane, Room 1061 Rockville , MD 20852

AMGEN

Paul R. Eisenberg, M.D., M.P.H. F.A.C.P., F.A.C.C. Senior Vice President

Global Regulatory Affa irs and Safety

Amgen Inc

One Arngen Center Onve Thousand Oaks, CA 91320 805.4474632 - Telephone 805.3757374 - Fax

Reference: Docket No. FDA·2011-D-0602, Draft Guidance for Industry on Quality Considerations in Demonstrating Bios imilarity to a Reference Protein Product; Availability

Reference: Docket No. FDA-2011-D-0605, Draft Guidance for Industry on Scientific Considerations in Demonstrating Biosimilarity to a Reference Product; Availability

Reference: Docket No. FDA-2011-D-0611 , Draft Guidance for Industry on Biosimilars: Questions and Answers Regarding Implementation of the Biologics Price Competition and Innovation Act of 2009

Dear sir or madam:

Amgen Inc. is pleased to respond to the FDA's draft guidance documents on the im plementation of the Biologics Price Competition and Innovation Act of 2009 (BPCI Act) as presented in the February 15, 2012 Federal Register Notices [Docket Nos. FDA-20 11-D-0602, FDA-2011-D-0605, and FDA-201 1-D-061 11 .

As one of the world 's leading innovative biotechnology companies and a future manufacturer of high quality biosimilar medicines, Amgen can offer a unique and valuable perspective on what is required to develop and manufacture safe and effective biolog ic drugs. Amgen 's innovative biologic therapeutics have helped millions of people around the world in the fight against cancer, kidney disease, rheumatoid arthritis , and other serious illnesses.

The attached document supplements Amgen 's 2010 comments on implementation of the biosimilar pathway and the testimony that was presented by Dr. Joseph Miletich at the November 3, 2010 public hearing. Amgen looks forward to participating in the May 11 , 2012 hearing on the draft guidances and recommendations for future biosimilar guidances.

Should you need any additional information regarding this submission, please contact Kimberly Greco by telephone at 202-585-9619 or by email at kgreco@amgen com

Sincerely,

Paul R. Eisenberg, M.D., M.P.H.

FDA Biosimilar Guidance Date: 16 April 2012 Amgen Response Page 1 of 42

Amgen, Inc AMGEN

Table of Contents

1. INTRODUCTION ........................................................................................................... 3

2. EXECUTIVE SUMMARY AND OVERVIEW ................................................................. 6

3. COMPLEXITIES OF PROTEIN PRODUCTS ............................................................... 9

3.1 Nature of Protein Products and Related Scientific Considerations .............................. 9

3.2 Manufacturing Process Considerations ...................................................................... 10

4. US –LICENSED REFERENCE PRODUCT AND OTHER COMPARATORS ........... 11

5. APPROACHES TO DEVELOPING AND ASSESSING EVIDENCE TO DEMONSTRATE BIOSIMILARITY ............................................................................ 15

5.1 Using a Stepwise Approach to Demonstrate Biosimilarity .......................................... 15

5.2 Using a Totality-of-the-Evidence Approach to Assess a Demonstration of Biosimilarity ................................................................................................................. 15

6. DEMONSTRATING BIOSIMILARITY ........................................................................ 16

6.1 General ....................................................................................................................... 16

6.2 Structural Analysis ...................................................................................................... 16

6.3 Animal Data ................................................................................................................. 17

6.4 Clinical Studies ............................................................................................................ 18

6.4.1 General ....................................................................................................................... 18

6.4.2 Human Pharmacology Data ........................................................................................ 18

6.4.3 Clinical Immunogenicity Assessment.......................................................................... 19

6.4.4 Clinical Safety and Effectiveness Data ....................................................................... 21

6.4.5 Extrapolation of Clinical Data Across Indications ....................................................... 25

6.4.6 Maintaining Biosimilarity ............................................................................................. 26

6.4.7 Class-Specific Guidances ........................................................................................... 26

7. POSTMARKET SAFETY MONITORING CONSIDERATIONS ................................. 28

7.1 Biosimilar Labeling ...................................................................................................... 28

7.2 The Need for Accurate Attribution of Adverse Events and Distinguishable Names ......................................................................................................................... 30

7.2.1 Special Considerations for Pharmacovigilance in the Biologics Context .................... 30

7.2.2 Suggested FDA Actions .............................................................................................. 33

8. OTHER ISSUES ......................................................................................................... 35

8.1 Publicly Available Information ..................................................................................... 35

8.2 Interchangeability ........................................................................................................ 36

8.3 Sample Retention ........................................................................................................ 38

8.4 Reference Product Exclusivity .................................................................................... 39


Amgen, Inc AMGEN

8.5 Definitions ................................................................................................................... 41

8.5.1 Protein ......................................................................................................................... 41

9. REFERENCES ............................................................................................................ 42


Amgen, Inc AMGEN

1. INTRODUCTION

Amgen Inc. (Amgen) welcomes the Food and Drug Administration’s (FDA’s) release in

February of three draft guidance documents on the regulation of biosimilars.1

As one of the world’s first biotechnology companies, Amgen has helped millions of

people around the world in the fight against cancer, kidney disease, rheumatoid arthritis,

and other serious illnesses, and are committed to putting patients first in everything we

do. Our comments on the draft biosimilar guidances are informed by Amgen’s 30 years

of experience discovering, developing and manufacturing biologics and our more recent

engagement in the process of developing biosimilars of non-Amgen products.

We

appreciate FDA’s commitment to working with stakeholders in a transparent guidance

development process to facilitate implementation of the Biologics Price Competition and

Innovation Act of 2009 (BPCIA).

Amgen’s longstanding support for a biosimilar pathway has been premised on three

fundamental principles that have also guided our evaluation of the Agency’s draft

Scientific, Quality, and Question and Answer Guidance documents. These are: patient

safety, sound science, and incentives for innovation. While we agree with much of the

content in these draft guidance documents, we believe certain changes are necessary to

advance patient safety and to promote confidence in biosimilars marketed in the United

States (US).

Amgen remains grounded in both our innovative and biosimilars development by a series

of scientific realities that – while sometimes demanding in terms of rigor, time and

resources – must underpin patient-focused biologic development programs and inform

the nature and extent of scientific evaluation. All biosimilars development programs will

grapple with some or all of these challenges and FDA should candidly communicate its

expectations in guidance.

I. Biosimilars are rDNA protein biologic medicines. In contrast to traditional

chemical drugs, rDNA protein biologics are very large molecules that often have

the potential to rely on multiple mechanisms of action enabled by one or more

1 “Scientific Considerations in Demonstrating Biosimilarity to a Reference Product” [Docket No. FDA-2011-D-0605] (“draft Scientific Considerations Guidance”); “Quality Considerations in Demonstrating Biosimilarity to a Reference Protein Product” [Docket No. FDA-2011-D-0602] (“draft Quality Considerations Guidance”); and “Q & As Regarding Implementation of the BPCI Act of 2009” [Docket No. FDA-2011-D-0611] (“draft Q&A Guidance”).


Amgen, Inc AMGEN

aspects of their structure. They are a heterogeneous mixture of related proteins,

rather than a homogenous collection of identical molecules; hence biosimilars are

neither expected nor required to be the same as the reference product. These

remarkable medicines facilitate important clinical benefits for patients, however,

the molecular size of these medicines makes them recognizable by a patient’s

immune system and thus they have a propensity to elicit an immune response

that could limit efficacy and/or present a safety concern in some patients. The

specific structural features of a molecule that cause immunogenicity are largely

unknown.

II. Biosimilars are not generic drugs. Biologics are typically much more complex

than chemical drugs and cannot be completely characterized even with today’s

scientific capabilities. Unlike generic drugs, biosimilars will be allowed to exhibit

some structural differences from their reference product in addition to using

different formulations, devices, and perhaps expression systems. However, to be

considered “highly similar”, the primary amino acid sequence of a proposed

biosimilar and its reference product should be the same.

III. Scientific understanding of reference products is often incomplete.

Analytics of protein biologics are improving; however, the understanding of the

relationship of molecular structure to clinical function may not be complete.

Increased understanding is possible through orthogonal interrogations and

biologic assays, however, both can be imprecise and leave factors unmeasured.

Although innovative manufacturers seek to characterize structural attributes that

are believed to be critical to safety and efficacy before products are licensed,

after licensure they regularly identify new critical quality attributes or combinations

of attributes that may affect the clinical profile.

IV. The “highly similar” standard for biosimilarity is burdened with uncertainty.

The uncertainty associated with biosimilarity can exist in two ways depending on

the approach a sponsor pursues. Where the sponsor purports to have matched

the attributes of the reference product, this approach is only as certain as the

knowledge of the reference product’s critical quality attributes is complete.

Alternatively, where a sponsor seeks to justify subtle differences between the


Amgen, Inc AMGEN

biosimilar and the reference product, the definition of an acceptable range of

differences that are not clinically meaningful is also uncertain.

V. Clinical testing of biosimilars will be needed for the foreseeable future. The

purpose of clinical testing is to demonstrate equivalent efficacy, safety, and

immunogenicity profiles of a biosimilar when investigated in parallel with the

reference product. Such clinical testing is neither intended nor designed to

establish de novo the complete safety or efficacy profile. Omission of

comparative clinical studies evaluating safety and efficacy will be appropriate only

if a biosimilar applicant can ensure, through other premarket testing, that there

are no clinically meaningful differences between the products’ clinical profiles. A

clinical evaluation of immunogenicity will always be essential.

VI. Accurate attribution of adverse events is essential to patient safety. Biologics are often highly sensitive to the changes in the manufacturing process

and the environment and structural alterations can result. Even minor structural

differences can have an important clinical effect on patients. FDA should ensure

effective pharmacovigilance for all protein products. In particular, the agency

should require measures that enable pharmacists, physicians, patients, health

authorities, and manufacturers to attribute adverse events to the appropriate

biologic (whether a biosimilar or a reference product). The agency should expect

manufacturers to employ multiple, redundant means of product identification.

Requiring biosimilars to bear distinguishable nonproprietary names is an efficient

means of facilitating accurate attribution of adverse events and product

traceability.

In the next section we provide an executive summary of our recommendations. In

subsequent sections we provide more detailed comments on the specific areas

addressed by the draft guidances. Amgen’s comments address the issues collectively,

although they appeared in more than one of the draft guidances. Thus, Amgen has

submitted a duplicate copy of these comments to each of the dockets established for the

three draft guidances.


Amgen, Inc AMGEN

2. EXECUTIVE SUMMARY AND OVERVIEW

Amgen commends the work of the FDA in drafting guidances that address many of the

issues important to implementation of the U.S. biosimilar approval pathway. With some

key modifications and additions, these will be useful resources for companies planning to

develop and manufacture biosimilars and for patients and physicians who will ultimately

determine the success of the new approval pathway. The following is an overview of our

recommendations for the three biosimilar draft guidances:

• Acknowledge what we know today and what we do not yet know. The science of

biotechnology has come a very long way in three decades but it would be

detrimental to patient safety and undermine the goal of achieving a successful

approval pathway if policy decisions were premised on aspirations rather than

achievements.

• Use more precise language in the final guidance documents, whenever feasible.

We recognize that regulatory standards and expectations must be flexible

enough to accommodate the uniqueness of each biologic and to adapt to

scientific and technical advances. The wide degree of flexibility and lack of clarity

reflected in the draft guidances, however, risk undermining the objective of

providing meaningful guidance to industry. Enhancing the specificity of the final

guidances would improve transparency and facilitate the timely development of

safe, pure, and potent biosimilars. We therefore recommend that the final

guidance documents address the following points:

Define ambiguous terms;

Discuss concepts with greater specificity (eg, more discussion of agency

expectations and examples); and

Acknowledge and address current technical and scientific limitations

related to regulatory advice or requirements, and expressly note the

challenges presented in evaluating the biosimilarity of proposed

biosimilars.

• Make clear that clinical studies will be necessary for the foreseeable future due to

the complexity and diversity of human biology and the limits of scientific

knowledge today. At the same time, it is appropriate for the agency to


Amgen, Inc AMGEN

acknowledge that analytical studies will increasingly facilitate reduction in the size

and scope of clinical studies needed.

• Permit a biosimilar sponsor to use data derived from a study conducted with a

foreign comparator to address components of statutory requirements if the

foreign comparator product can be determined to be fully representative of the

US licensed reference product. As the information relevant to demonstrating this

may be held confidential by the reference product sponsor, FDA should specify

that foreign comparator product data can be used if two criteria are satisfied: first,

a single company manufactures and distributes both the US licensed and foreign

comparator product and, second, the biosimilar applicant establishes a scientific

basis for relying on comparative data involving the foreign product.

• Acknowledge biosimilars as stand-alone products for purposes of ongoing

regulation once they have been approved as safe and effective. A requirement to

maintain biosimilarity between the biosimilar and the reference product would

unnecessarily complicate the regulatory process without providing meaningful

benefit to patients and be likely to increase both regulatory and manufacturing

costs.

• Commit to publishing class-specific guidance that discusses the approval

standards and other key implementation issues for particular product classes.

Class-specific guidance will provide the most meaningful guidance for

manufacturers and will foster the confidence in biosimilars needed for physician

and patient acceptance of this new class of product.

• Require all biological products to have a distinguishable nonproprietary name in

order to facilitate accurate attribution of adverse events. The distinguishable

name could consist of a common root and unique suffix, such as a Greek letter or

the manufacturer’s name, to enable identification of the product and manufacturer

as well as to provide clear class affiliation.

• Specify that biosimilar product labeling must provide all information necessary for

physicians and patients to make informed choices. This includes a standard

biosimilar label format as well as key biosimilar and reference product data,

clearly labeled as such.


Amgen, Inc AMGEN

• Make clear that the agency is committed and fully intends to respect and enforce

protections against disclosure of trade secret and commercial confidential

information.

• Address the key issues of drift and attribution of adverse events following product

switching before designating any product as interchangeable. Interchangeability

is a higher standard of approval than biosimilarity and presents a complex set of

regulatory challenges that should be addressed fully in a separate guidance.

• Adopt separate regulations, via notice and comment rulemaking, regarding the

retention of reserve samples rather than simply “recommending” reliance on the

long-standing requirements for ANDA sponsors. Unique considerations relating

to recombinant biological products make this necessary.

• Clearly acknowledge that reference product exclusivity is presumptively granted

by the law and the requirement for proof arises only to the extent a product could

be considered a subsequent generation product. It would be contrary to the plain

language of the statute to shift the burden of proof to the applicant to “make its

case” for exclusivity when the applicant files a stand-alone BLA filed under

351(a). Uncertainty about whether a product will be eligible for data protection

will severely impede biotech development and will significantly deter investments

that improve products and therefore medical advances for patients.


Amgen, Inc AMGEN

3. COMPLEXITIES OF PROTEIN PRODUCTS

3.1 Nature of Protein Products and Related Scientific Considerations

As FDA correctly notes in the draft Scientific Considerations Guidance (lines 156 - 158),

“analytical methodologies have increasingly improved the ability to identify and

characterize not only the drug substance of a protein product, but also excipients and

product- and process-related impurities.” The draft also correctly acknowledges that

“current analytical methodology may not be able to detect all relevant structural and

functional differences between two proteins” (lines 160 - 162). We believe it is critically

important that FDA acknowledge that the inherent uncertainties and challenges

associated with demonstrating biosimilarity are not purely – or even largely – related to

the sensitivity and specificity of characterization methods. Rather, one of the primary

challenges in assessing biosimilarity is knowing which structural attributes contribute

meaningfully to product safety and efficacy.

The fundamental assumption underlying reliance upon analytical and nonclinical

comparison must be that the sponsor and regulator know the precise relationship

between structural attributes and the clinical performance of the medicine and have the

ability to design the appropriate assays to mimic all in vivo effects in humans. Today,

this is a flawed assumption in terms of predicting the efficacy, safety and immunogenicity

of protein medicines. While our collective knowledge of the structure-function

relationship may well improve over time – perhaps even through biosimilars development

– we need to be candid about the current state of science while not excluding future

development.

The Agency states that “data derived from analytical studies, animal studies, and a

clinical study or studies are required to demonstrate biosimilarity unless FDA determines

an element unnecessary,” (lines 162 – 164) but the guidance documents offer few

insights into the circumstances where these waivers may be justified. Amgen believes

FDA should clarify that waivers should be sought only in those circumstances where

equivalent efficacy and non-inferior safety and immunogenicity can be assured through

other means before approval. Given that even seemingly minor modifications may exert

unintended clinical effects – especially given a protein’s propensity to change

conformation in vivo and elicit unexpected immunological events – Amgen cannot

contemplate circumstances in which clinical studies would not be needed today and in

the foreseeable future in order to assure an absence of clinically meaningful differences.


Amgen, Inc AMGEN

3.2 Manufacturing Process Considerations

We agree with the statement in the draft Scientific Considerations Guidance that

“demonstrating that a proposed product is biosimilar to a reference product typically will

be more complex than assessing the comparability of a product before and after

manufacturing changes made by the same manufacturer” (lines 181-183). As both an

innovative product sponsor and a biosimilar developer, Amgen has extensive knowledge

of the differences between these two scientific exercises. These include differences in

the sponsor’s clinical experience concerning the relevant product, knowledge of the

structure-function relationship of the product, and access to historical data for the

product.

Because of changes an innovative product sponsor may make to its manufacturing

process, a reference product studied in a biosimilar development program may change

(either during or after product development) in subtle ways that are unknown to the

biosimilar applicant. To address this concern, FDA should recommend that a biosimilar

applicant conduct its analytical, nonclinical, and clinical comparisons during a limited

period of time and then establish an internal reference standard. Neither the biosimilar

nor the reference product should be constrained from improving their processes and

complying with cGMP following approval of a biosimilar product. The biosimilar and

reference products should be treated as independent products, each standing on their

own merits for future manufacturing changes and quality evaluations. To do otherwise

would unduly constrain manufacturing improvements across the industry. The sensitivity

of biologics to the manufacturing process is an issue FDA will have to address prior to

designating one biologic product to be interchangeable with another. Additional

considerations related to changes in attributes of the proposed biosimilar or the

reference product due to manufacturing changes are discussed in greater detail in

Section 6.5.6, below.


Amgen, Inc AMGEN

4. US –LICENSED REFERENCE PRODUCT AND OTHER COMPARATORS

The draft Scientific Considerations Guidance states that “[t]o obtain licensure of a

proposed product under section 351(k) of the Public Health Service (PHS) Act, a

sponsor must demonstrate that the proposed product is biosimilar to a single reference

product that previously has been licensed by FDA… However, under certain

circumstances, a sponsor may seek to use data derived from animal or clinical studies

comparing a proposed product with a non-U.S.-licensed product to address, in part, the

requirements under section 351(k)(2)(A) of the PHS Act” (lines 197 - 199, 204 - 206)

(emphasis added). See also draft Q&A Guidance § A.I.8. Where a foreign comparator

product can be determined to be fully representative of the US licensed reference

product, Amgen believes that a sponsor should be allowed to use data derived from a

study conducted with this foreign comparator to address components of the statutory

requirements. However, the assessment of whether a foreign comparator product can

be shown to be fully representative of the US licensed reference product is – in and of

itself – a complex challenge.

• The challenge - Foreign versions of US-licensed reference products can significantly

vary in the extent to which they are similar to the US product. In some cases, a

single manufacturer may release the same drug product (with differences only in

packaging and labeling) from the same manufacturing plant(s) into both the US and

ex-US markets. In other cases, products intended for different markets may be

manufactured independently or comply with locally-specific quality or pharmacopeial

standards. Where products are manufactured by more than one entity, two products

that were once related may have diverged from each other over time in subtle but

important ways. Moreover, some innovative products manufactured abroad may

have even more fundamental differences from the products intended for marketing in

the US, for example due to the use of different cell lines, formulations, container

closure, or shipping methods. In light of the many ways in which foreign versions of

US-licensed reference products can resemble or differ from the US version of the

product, Amgen urges FDA to take a highly cautious approach to relying on

comparative data involving a foreign comparator product to support the

demonstration of biosimilarity to a US-licensed reference product.


Amgen, Inc AMGEN

• The standard - Amgen believes it is scientifically appropriate, and consistent with

FDA’s authority under the statute, to permit reliance on a foreign comparator product

when it is the same as the US licensed product but for regionally-specific packaging

or label (ie, fully representative of the product US patients currently receive including

use of the same nonclinical and clinical data supporting licensure). We recognize

that this information may be held as trade secret or confidential commercial

information and believe protection of this information is both appropriate and

important. Therefore, we suggest that the final biosimilar guidance specify two

criteria that biosimilar applicants must address to demonstrate that the foreign

comparator product is fully representative of the US-licensed reference product.

• The proposed policy - First, if available evidence indicates that a single innovator

company manufactures (directly or by a subsidiary or contract manufacturing

organization under the supervision of the BLA holder’s quality unit) the innovative

product and distributes it globally (including in the US), it is possible that the foreign

comparator product purchased outside the US could be fully representative of the

US-licensed reference product. Thus, reliance on comparative data involving this

foreign product may in some cases be scientifically appropriate, if the applicant

provides adequate scientific justification (as discussed in the second criteria below).

• In contrast, if a foreign product were manufactured by an entity other than the US

license holder, it would be unlikely that the foreign product could be fully

representative of the US-licensed product. Therefore, data concerning this product

should not be used to demonstrate biosimilarity.

• Second, even if the foreign product is manufactured by the US-license holder or

supervised subsidiary, the applicant must bear the burden of establishing a scientific

basis for relying on comparative data involving the foreign product given the

attendant potential for regionally distinct products. We suggest that FDA require the

applicant to make the following showings, at a minimum, to permit reliance on data

involving a foreign comparator product:

1. The biosimilar applicant must be able to fully characterize the two innovative

products and execute the testing necessary to detect differences between the

products, including designing and executing appropriate assays and studies and

appropriately identifying the limitations of the sampling and analytical


Amgen, Inc AMGEN

methodology. The analytical comparison bridging the products must be

representative of products used in the clinical confirmatory study. This should not

be an exercise of historical controls or reference but actual side-by-side testing

relevant to the pertinent confirmatory clinical study.

2. The manufacturing facilities for the foreign comparator product are in compliance

with current good manufacturing practices (cGMPs) and are licensed by and

subject to regularly scheduled inspections by FDA or equivalent regulatory

agency. This showing should be made with respect to both the drug substance

and the drug product. If information about the specific facility in which the foreign

comparator product is manufactured is unavailable, we believe that ensuring that

the two innovative products have the same license holder may in some cases

serve as a surrogate, in conjunction with the additional factors below, for showing

compliance with cGMPs and regularly scheduled inspections for the specific

facility in which the foreign comparator product was manufactured.

3. Chemical, analytical, pharmacology, and clinical pharmacokinetic (PK) data

demonstrate that the foreign comparator product and US-licensed reference

product – both drug product and drug substance, where feasible – are fully

representative of each other (ie, structural properties and quality attributes are

practically indistinguishable other than differences attributed to the variability of

the test methods or lot-to-lot variability of the US-licensed reference product).

4. The foreign comparator product has the same formulation as the US-licensed

reference product.

5. The foreign comparator product has the same container closure system as the

US reference product, or if differences exist the applicant bears the burden of

showing that the reference drug product and foreign comparator are the same

irrespective of any difference in the container closure system.

6. The foreign comparator product is approved and has been widely marketed for

an appropriate length of time in a region with a similarly robust, International

Conference on Harmonisation (ICH) regulatory regime comparable to FDA. In

particular, FDA should seek assurance that the relevant regulatory authority has

adverse event reporting requirements and a postmarket surveillance system at

least as robust as those in the US.


Amgen, Inc AMGEN

Amgen believes that if a biosimilar applicant makes these showings with respect to a

foreign comparator product and provides adequate scientific justification, it will be

appropriate for FDA to consider accepting for review the data obtained from comparing

the proposed biosimilar to the foreign comparator product.

In summary, if a reference biological product is manufactured by the US license holder

for distribution in multiple regions such that a foreign comparator product can be shown

to be fully representative of the US reference product by the above criteria, then the

agency may consider comparative nonclinical or clinical studies using the foreign

comparator product to be supportive of a demonstration of biosimilarity. In these limited

circumstances it would be reasonable for FDA to consider permitting the use of such

comparative studies in lieu of repeating nonclinical or clinical studies using the US-

licensed reference product.


Amgen, Inc AMGEN

5. APPROACHES TO DEVELOPING AND ASSESSING EVIDENCE TO DEMONSTRATE BIOSIMILARITY

5.1 Using a Stepwise Approach to Demonstrate Biosimilarity

An assessment of biosimilarity based on a stepwise comparison of the proposed

biosimilar and the US-licensed reference product must include certain basic

components. We believe that a biosimilar development program should include, at a

minimum, side-by-side analytical evaluations as well as comparative nonclinical studies

(in vitro and/or in vivo), comparative human PK and pharmacodynamic (PD) studies

(when relevant response-marker exists), and a comparative clinical study addressing

safety and efficacy separately or jointly.

5.2 Using a Totality-of-the-Evidence Approach to Assess a Demonstration of Biosimilarity

The concept of a totality-of-the-evidence approach can serve as an appropriate and

efficient approach to demonstrating biosimilarity. FDA should adopt the following

specific principles, however, to ensure that a biosimilar development program meets the

statutory standards for approval:

• Studies should be conducted with appropriate temporal boundaries. The

assessments should use state-of-the-art methods, consistent test material and

reference product, consistent reagents, and sensitive assays.

• Safety and immunogenicity should be evaluated in appropriate preapproval

clinical studies and, where appropriate, postmarketing studies of sufficient

duration to detect and assess differences in a sensitive population. Population

models or other assumptions should be used only when validated by applicant-

derived data concerning both the reference product and the proposed biosimilar.

• Biosimilar applicants and the agency should take robust and transparent

measures to ensure that trade secret and confidential commercial information is

appropriately protected.


Amgen, Inc AMGEN

6. DEMONSTRATING BIOSIMILARITY

6.1 General

A biosimilar product must have an identical amino acid sequence and must be highly

similar to the reference product in multiple higher order structural and functional

attributes. This is critical because differences in protein folding, cross-linking, or subunit

association patterns could affect the functionality, immunogenicity, stability, and/or PK of

the biological product. Given the highly complex and heterogeneous nature of biologics,

existence of lot-lot variation, and distinctive degradation profiles, Amgen recommends

that FDA articulate a clear, statistically rigorous model for a biosimilar sponsor to sample

and evaluate the appropriate number of lots of the reference product and proposed

biosimilar product while ensuring a representative sampling across the shelf life.

Sampling plans may well need to account for variability on an attribute by attribute basis.

6.2 Structural Analysis

The draft Scientific Considerations Guidance (lines 323 - 325) and draft Quality

Considerations Guidance (lines 369 - 371) imply that in some circumstances, the amino-

acid sequence of a biosimilar and its reference product may be different. We

recommend that, at a minimum, the amino acid sequence (primary structure) of the

biosimilar must be identical to that of the reference product, notwithstanding minor

differences in the extent of N-terminal or C-terminal processing. Neither by negligence

nor design should primary structure differences exist. Other minor sequence variants

comprising low prevalence gene sequence mutations or errors in translation are

sometimes evident and their impact on safety and efficacy must be assessed. FDA

should clarify the extent to which sequence variants exist within the concept of

biosimilarity. Not only is this appropriate as a scientific matter, but we believe the BPCIA

prohibits virtually all differences in amino acid sequence.

Section 351(k)(2)(A)(i)(I)(aa) requires that a biosimilar application be supported by data

from “analytical studies that demonstrate that the biologic is highly similar to the

reference product notwithstanding minor differences in clinically inactive components”

(unless waived by FDA). We believe that a difference in amino-acid sequence beyond

terminal truncations would necessarily preclude a finding that two products were “highly

similar” and that such a difference would not constitute “minor differences in clinically

inactive components.” Further, the applicant should bear the burden of conclusively


Amgen, Inc AMGEN

establishing that terminal truncations have no impact on efficacy, safety and

immunogenicity.

6.3 Animal Data

Amgen agrees that a nonclinical in vivo safety evaluation of a biosimilar should generally

be conducted before clinical studies. We suggest that FDA provide greater clarity with

regard to when animal toxicity studies will be appropriate to help inform biosimilar

development program planning. We believe it would also be helpful for FDA to address

this issue in class-specific guidance.

However, Amgen suggests strengthening the statement in the draft Scientific

Considerations Guidance that “animal toxicity studies are generally not useful if there is

no animal species that can provide pharmacologically relevant data” (lines 404 - 405).

We suggest that the final guidance provide that toxicity studies in a nonrelevant species

are not warranted and should not be conducted. Additionally, since some monoclonal

antibody product sponsors have historically used the chimpanzee as the only

pharmacologically relevant nonclinical species, we ask that FDA clarify that it

discourages the use of chimpanzees for the evaluation of biosimilars for ethical reasons.

In support of replacing, refining, and reducing animal testing (3Rs), FDA should clarify

the minimum expected study design for animal toxicity studies. The PHS Act states that

an application must include information demonstrating that the biologic is biosimilar to a

reference product based upon data derived from animal studies (including the

assessment of toxicity), suggesting that the toxicology study should be designed to

compare the two products (not re-establish the toxicity profile of the reference product).

Accordingly, the final guidance should address whether a toxicity study design may use

one dose level, may use one gender, may omit a recovery group, or may omit a vehicle

control group. For example, Amgen suggests that evaluation of a single gender in the

toxicity study may be sufficient to characterize the biosimilar if the more sensitive gender

is selected, if there are no gender differences in effects, or if the reference product is

primarily used in only one gender (eg, trastuzumab [Herceptin®]). Also, in the case of

biologics for which more than one species is pharmacologically relevant, Amgen believes

it would be sufficient to evaluate a single species. If more than one species is equally

sensitive, the lowest order species should be used.


Amgen, Inc AMGEN

Amgen agrees that nonclinical safety pharmacology, reproductive and developmental

toxicity, and carcinogenicity studies usually will not be warranted. Because ICH

Guideline S6(R1) states that tissue cross-reactivity studies are not recommended for

assessing comparability of the test article as a result of process changes, Amgen

suggests noting that evaluating “tissue cross-reactivity” also is not warranted.

6.4 Clinical Studies

6.4.1 General

Amgen recommends that the final guidance documents discuss with greater specificity

FDA’s expectations about the nature and extent of clinical studies needed to

demonstrate biosimilarity. We suggest that applicants be expected to provide, at a

minimum, comparative PK data, comparative PD data (where relevant dynamic markers

exist), comparative immunogenicity data, and comparative efficacy and safety data in

one or more conditions of use for which the reference product is licensed. The nature of

the proposed biosimilar product or its class may in some cases permit an applicant to

deviate from this general expectation.

6.4.2 Human Pharmacology Data

Amgen agrees with the recommendation in the draft Scientific Considerations Guidance

that an applicant should use a population PK modelling approach to explore potential

variability in PK (lines 700 – 707). If, as expected, only a few clinical studies evaluating

the proposed biosimilar will be conducted, it would not be possible to conduct an

extensive evaluation of the covariates in a population PK model, as is recommended in

FDA guidance. See FDA Population Pharmacokinetics Guidance, 1999. Any use of

population PK models should be scientifically justified by the applicant on the basis of

current, side-by-side scientific data using state-of-the-art methods.

The draft Scientific Considerations Guidance also provides that establishing a similar

human PK and PD profile may provide a basis for a selective and targeted approach to

subsequent clinical testing (lines 516 - 518). We suggest that the final guidance more

clearly explain the expected role of PK and PD data. Although the real value of PK and

PD data will depend on the product class and data generated by the biosimilar applicant,

the final guidance should nevertheless establish general expectations about the nature


Amgen, Inc AMGEN

and extent of clinical studies beyond PK and PD data that will be required to support

biosimilar approval.

6.4.3 Clinical Immunogenicity Assessment

The draft Scientific Considerations Guidance appropriately emphasizes the importance

of clinical immunogenicity assessments to establishing biosimilarity (lines 540 - 542).

Post-translational modifications are frequently dependent on manufacturing process and

therefore could contribute to an alteration in the immunogenicity of the biosimilar

compared to the reference product. Even advanced analytical techniques may not

detect differences between proteins that could affect the rate or nature of immunogenic

reactions. Predictive immunogenicity models are promising but not currently validated.

In addition, nonclinical models do not reliably predict immunogenicity. Clinical studies

are therefore necessary to assess immunogenicity, and the resulting data in turn are

necessary for the assessment of biosimilarity. Because anti-drug antibodies (ADA) may

affect PK, PD, and product safety parameters, their potential effect should be evaluated

in a sufficiently powered preapproval study.

We suggest that FDA define “immunogenicity” and “immune response” and make clear

that they are not synonymous. The draft Scientific Considerations Guidance, for

example, inappropriately suggests that a “cytokine response” may be due to

immunogenicity (see lines 584 - 587) rather than an immune response. We further

recommend that the term “immunogenicity” be used to refer to the potential of a product

to generate ADA. We recommend that the final guidance documents use the phrase

“immune response” to refer to immune reactions such as cytokine release or

hypersensitivity reactions. Hypersensitivity reactions, for example, can occur

spontaneously and/or indicate the presence of a pre-existing specific or cross-reacting

antibody or a nonspecific reaction.

The draft Scientific Considerations Guidance states that if immune response to the

reference product is rare, two studies may be needed to evaluate immunogenicity: (1) a

premarket study to detect major differences in responses; and (2) a postmarket study to

detect subtle differences (lines 555 - 559). We suggest that FDA explain what types of

differences in ADA generally would constitute a major difference. Additional guidance on

the types of differences in ADA that would constitute a major difference could be

provided in class-specific guidances.


Amgen, Inc AMGEN

We agree with the draft Scientific Considerations Guidance’s recommended use of a

comparative parallel design (ie, a head-to-head) study to assess potential differences in

immunogenicity risk (lines 562 - 564). We also agree that the products should be

assessed in the same assay with the same patient sera whenever possible (lines 612 -

613). Even if a biosimilar applicant attempts to standardize the assays, cross-study ADA

incidence rates cannot be fully compared between different assays due to the

dependence of detection on both antibody affinity and concentration. Analytical

advances have improved immunogenicity assay performance over time, making it

difficult to compare immunogenicity incidence rates measured using different methods.

Therefore, it is essential for an applicant to test both products in the same assay at the

same time.

The draft Scientific Considerations Guidance states that the selection of clinical

immunogenicity endpoints or PD measures should take into consideration

immunogenicity issues associated with the reference product (lines 584 - 590). For

some innovative products, however, the incidence of immunogenicity, or the correlation

between immunogenicity and clinical events, may not be well known. We suggest that

FDA provide additional information about the endpoints to be used in these cases, as

well as examples of criteria for determining acceptable immunogenicity endpoints.

A follow-up period (lines 592 - 599) is important to appropriately characterize the time

course of immunogenicity (eg, as pre-existing or developing after treatment). We

recommend that subjects positive for ADA at the end of the study be followed for at least

one year after the last administration of either the biosimilar or the reference product.

We also recommend that it state that the sampling frequency for chronically administered

products should be at least every three months during the study and follow-up periods.

We agree that characterization of the incidence and severity of ADA is necessary to

understand immunogenicity elicited by both products (Scientific Considerations guidance

lines 535 - 537). We recommend that FDA explain when such characterization is

necessary. Specifically, we recommend that subjects who are positive for ADA be

characterized for antibody titer when a clinical effect (such as loss of efficacy) has been

identified. The Agency should note that specificity characterization is needed when the

proposed product contains multiple domains that may have differential effects. Further, it

should state that isotype characterization will be expected when transient ADA (possible

IgM) or hypersensitivity reactions (possible IgE) are observed.


Amgen, Inc AMGEN

We agree with the statement in the draft Scientific Considerations Guidance that to

permit extrapolation of immunogenicity data from evaluation of the proposed biosimilar

product in one indication to support the product’s licensure for another, a sensitive

population should be studied (lines 576 - 582). At a minimum, the population studied

should be immunocompetent (if there is such a population for the approved use).

6.4.4 Clinical Safety and Effectiveness Data

General Considerations

The draft Scientific Considerations Guidance states that comparative clinical safety and

effectiveness data are necessary if there are “residual uncertainties” about biosimilarity

based on structural and functional characterization, animal testing, human PK and PD

data, and a clinical immunogenicity assessment (lines 620 - 623). We believe that at

least one clinical study evaluating safety and effectiveness will be needed in every case

to establish an absence of clinically meaningful differences between the reference

product and the proposed biosimilar. A biosimilar development program that relies on

only analytical and nonclinical comparisons would depend on the assumption that the

biosimilar sponsor knows all structural attributes that may affect the clinical performance

of the reference product and the proposed biosimilar. It would also depend on the

assumption that the biosimilar sponsor is able to design appropriate nonclinical assays

that identify all aspects of the product’s clinical profile.

Given the current state of science, these assumptions are not warranted and they raise

patient safety risks including that the molecule may have different efficacy. Even

seemingly minor modifications in the structure of a biologic, its manufacturing process, or

other product characteristics may result in unintended clinical effects. This is particularly

true given a protein’s propensity to change conformation in vivo and to elicit unexpected

immunological events. Although the general scientific community’s understanding of the

relationship between a particular protein’s structure and its function may improve over

time, this potential for future scientific advancement does not justify the use of

inappropriately abbreviated biosimilar development programs today. Accordingly, given

current limitations of the science, FDA should require at least one clinical study

evaluating safety and effectiveness, including immunogenicity.


Amgen, Inc AMGEN

Statistical Considerations

Equivalence should be the base expectation for clinical efficacy studies unless the

applicant can justify the use of a one-sided non-inferiority (NI) test. We note that FDA

highlights the fact that an NI design could be advantageous as it would allow for a

smaller sample size than an equivalence design. We believe that it is important for FDA

to remove the term “advantageous” as for fixed margin designs; the sample size required

for an NI design (1-sided alpha level of 0.025 and with at least 80% power) is no more

than 10% less than that of an equivalence design. The advantage in terms of sample

size is therefore relatively minor and so it should not be stated by FDA as an advantage

and nor should the smaller sample size be positioned as one that has any bearing on

decision making when it comes to study design. In the event that NI is considered, the

applicant should address not only the efficacy considerations but also the potential for

increased toxicity that may be associated with higher potency/effect that is not formally

evaluated in an NI study.

It is stated clearly in the guidance that clinical studies should be designed such that they

can demonstrate the proposed product has neither decreased nor increased activity

compared to the reference product, (ie, they should be equivalence trials). If the

proposed product has superior efficacy than the reference product, then it must be

treated as a distinct product and follow the appropriate licensure pathway section 351(a)

of the PHS act. The circumstances where a noninferiority study for efficacy may be

scientifically appropriate, as discussed in the guidance, seem to be rare. As a practical

matter, it is unlikely that such a study can provide sufficient evidence to meet the

statutory requirement for showing an absence of clinically meaningful differences. A

noninferiority assessment leaves open more uncertainty and hence extrapolation to other

indications with this increased uncertainty would be inappropriate. Therefore, we

recommend a showing of equivalent efficacy as a prerequisite for biosimilarity.

For equivalence trials, the draft Scientific Considerations Guidance states that the most

straightforward study design is a two-sided test in which the null hypothesis is that either

(1) the proposed product is inferior to the reference product or (2) the proposed product

is superior to the reference product based on pre-specified equivalence margins (lines

665 – 669). The guidance further suggests that a different upper and lower margin may

be appropriate. We agree with the Agency’s consideration in maintaining equivalence in


Amgen, Inc AMGEN

principle, but with potential flexibility on the superiority side. Along the same line of

thinking, we suggest considering asymmetric Type 1 errors.

Criteria for biosimilarity are not defined in the guidance for any of the steps in the

stepwise approach or for the totality-of-the-evidence. Five example factors listed in

VII.D.3 are listed for determination of the equivalence margins in the comparative clinical

studies. We agree and believe that the characteristics, including an allowance for

accurately assessed variability of the reference product (ie, not due to limited nature or

other insufficiency of sampling program) are essential in setting the criteria for

biosimilarity (and/or interchangeability) in all the steps in the stepwise approach and for

the totality-of-the-evidence. The draft guidance does not address the appropriate

number of lots for adequate sampling to assess lot-lot variability of the biosimilar and

reference product, nor the appropriate sourcing model for clinical trials.

Other Design Considerations

We believe that biosimilar applicants should be required to provide data from a

randomized head-to-head clinical study(s) (confirming clinical biosimilarity with the US-

licensed reference product) in a sensitive population to demonstrate biosimilarity. The

Agency should note that in some cases evaluation of more than one sensitive population

may be necessary. In addition, an applicant should be required to justify the clinical

endpoints used. We believe that although the endpoints may be the same as those

studied by the innovator, surrogate endpoints may be considered on a case-by-case

basis. We also recommend that the clinical trials be powered to demonstrate equivalent

efficacy and (even if not powered for statistical significance) should be able to describe

noninferior safety outcomes.

In addition to specific safety and efficacy endpoints, PK studies should be required

before licensure of all biologics. Because rare and important immunologic outcomes are

sometimes not observed in the prelicensure setting we acknowledge the limitation of

preapproval data particularly in the event of approved indications without clinical data

(extrapolated indications). And while we support continued and careful analysis of

immunologic parameters within the sponsors’ clinical study program, we emphasize the

importance of careful pharmacovigilance in the postmarketing environment allowing

traceability of any adverse event to the level of the product, manufacturer and lot of the


Amgen, Inc AMGEN

medicine the patient received. The rationale for this is demonstrated by Europe’s

experience with pure red cell aplasia (PRCA), as discussed in Section 7.2.1below.

Safety considerations

The draft Scientific Considerations Guidance describes FDA’s expectations with regard

to the analysis and comparison of the efficacy and immunogenicity profiles of the

proposed biosimilar and its reference product. We recommend that FDA also describe

how data concerning the safety profiles of the products should be assembled and

compared and how a biosimilar applicant should demonstrate that there are no clinically

meaningful differences between the products’ safety profiles. FDA should clarify how an

applicant may demonstrate the similarity of both: (1) the incidence rates of “common”

adverse events that may be nonspecific (eg, injection site reactions) or mechanism

based (eg, dermatologic reactions to epidermal growth factor receptor antibodies); and

(2) rare serious adverse events (eg, lymphoma with tumor necrosis factor inhibitors) that

may or may not be mechanism-based. Comparing product safety profiles with regard to

rare serious adverse events, for example, would be extremely challenging; a very large

or long-duration study would be anticipated to detect a single event, let alone

differentiate between the adverse event profiles of two products.

We believe that although clinical studies will not necessarily be designed or powered to

detect statistically significant differences in safety profiles, a proposed biosimilar’s

preapproval safety database should be sufficiently robust to exclude important

differences in known adverse events associated with the reference product. It will be

possible for an applicant to determine a proposed biosimilar’s risk for common adverse

events relative to that of the reference product based on the historically reported

incidence rates of adverse events associated with the reference product. A sample size

calculation could then determine detectable differences between the safety profiles of the

two products.

As an example, Table 1 below demonstrates how such a calculation could be conducted

with a granulocyte-colony stimulating factor (G-CSF) biosimilar. Table 1 describes the

number of study subjects needed to determine whether a proposed biosimilar G-CSF

product is associated with increased risk for adverse events commonly reported in

G-CSF clinical trials (where the risk is a 2-, 3-, or 4- fold increase compared to that of the

reference product). In general, we would consider a 2-fold increase of risk (relative risk =


Amgen, Inc AMGEN

2) clinically meaningful for most adverse events. Of the events reported in Table 1, the

most clinically important are bone and musculoskeletal pain. These are the events that

should determine the size of the safety database. In this case, about 150 patients

treated with the biosimilar G-CSF (and 150 patients treated with the reference product)

would be needed to exclude a 2-fold increase in incidence of bone and musculoskeletal

pain. This sample size would thus enable reasonable conclusions to be drawn about the

similarity of the general (ie, common adverse event) safety profiles of the two products.

Table 1. Sample Size Estimation for Different Background Rate and Relative Risk Based on Observed Incidence Rates of Common AEs from G-CSF Clinical Study

Adverse Events Reported Incidence Rate for G-CSF (%)

Sample Size in Each Study Arm for Various Relative Risk Level

RR=2.0 RR=3.0 RR=4.0

Bone Pain 30 141 55 35

Abnormal Labs 3 829 281 159

Nervous System Events 16 199 72 43

Headache or Respiratory

8 341 119 69

Tiredness or Chest Pain 6 438 151 87

Musculoskeletal Pain 44 133 55 36 Note: All estimations are based on two side test, α = 0.05, Power (1-β) = 80%

The biosimilar’s labeling should indicate the 95% confidence limits for the point estimates

of rates of adverse events observed in clinical trials. If the difference in rates of adverse

events is notable but the biosimilarity standard is nevertheless deemed satisfied,

however, the labeling should identify the fold increase in risk of adverse events excluded

at a 95% confidence level.

A similar approach could be taken for other products or classes.

6.4.5 Extrapolation of Clinical Data Across Indications

Amgen agrees that a biosimilar applicant should scientifically justify the extrapolation of

clinical data to support a determination of biosimilarity for each condition of use of the

reference product for which licensure is sought. Draft Scientific Considerations

Guidance lines 765 - 767 and draft Q&A Guidance § A.I.11. We agree that clinical data


Amgen, Inc AMGEN

may be extrapolated to support licensure of other indications only if the data were

sufficient to demonstrate the safety, purity, and potency of the biosimilar in the indication

studied clinically. Draft Scientific Considerations Guidance lines 760 -7 64 and draft

Q&A Guidance § A.I.11. Extrapolation would be inappropriate without this foundation.

We also support the list of factors in the draft Scientific Considerations Guidance (lines

769 - 795) that the scientific justification for extrapolation should address.

6.4.6 Maintaining Biosimilarity

The draft guidance documents do not address considerations related to demonstrating

that biosimilarity is maintained after approval. We recommend that the final guidance

documents note that there is no statutory requirement that biosimilarity be maintained

after licensure and that neither the biosimilar sponsor nor the reference product sponsor

is constrained from improving its manufacturing process in compliance with cGMPs.

Requiring the maintenance of biosimilarity over time would inhibit manufacturing and

quality improvements and unduly burden industry without benefiting patients. Amgen

believes that FDA’s finding of biosimilarity at a specific point in time provides a sufficient

basis for the products to be considered as “stand alone” products thereafter. Biosimilar

applicants should be expected to establish an internal reference standard to facilitate all

postapproval comparability exercises performed by a biosimilar sponsor before and after

any manufacturing change that it makes.

6.4.7 Class-Specific Guidances

We strongly encourage FDA to develop, on a timely basis, class-specific biosimilars

guidance, akin to those developed by the European Medicines Agency. We believe FDA

should prioritize classes that contain products near patent expiry. Class-specific

guidances are vital to transparency. By providing greater specificity about regulatory

standards and expectations than do the draft overarching guidance documents, class-

specific guidance will promote more productive interaction between applicants and

reviewers. They will help conserve biosimilar applicant resources and agency review

resources and facilitate consistent treatment of biosimilar applicants. In addition, the

process of developing these guidances will itself promote public understanding of the

approval process, physician education, and confidence in licensed biosimilars.


Amgen, Inc AMGEN

Specifically, we recommend that the regulatory expectations for technical areas such as

the following topics be defined by the Agency (in consultation with industry and the

scientific community) in product class specific guidances:

1. Define the nonclinical studies that would be considered necessary to precede

clinical studies. This is of particular value in the context of defining the nature

and extent of animal studies and potentially whether such studies, for a precise

product/class would be necessary.

2. Define the endpoints for PK and PD studies recommended by the agency and

provide guidance on the design of such studies. This is of particular value in

defining the ultimate value of PD studies in the context of supporting the

establishment of equivalent efficacy to that of the reference product.

3. Based on the current level of scientific understanding, a biosimilar sponsor will

need to establish equivalent efficacy and noninferior safety in a formal clinical

study(s), beyond PK and PD. Define the patient population that is considered

most sensitive for detecting clinically meaningful differences in safety, efficacy

and immunogenicity. In some cases this can be one population, in other cases

two or more populations may need to be considered.

4. Define the design considerations for the safety/efficacy study(s), including the

endpoint(s) to be evaluated, duration of treatment, duration of follow-up and

statistical consideration for establishing biosimilarity.

5. In terms of safety, define the safety events of interest where differences in

severity or frequency may be considered clinically meaningful.

6. In terms of immunogenicity, address the patient population, duration of treatment

and antibody testing strategy that would be necessary to provide evidence of no

worse immunogenicity.

7. Finally, as the pre-approval studies are by nature limited and without prejudice to

approval of a biosimilar by the Agency, define clinical events to be followed

during formal post approval commitments, including the precise clinical events

and suggestion of how they would be evaluated (passive pharmacovigilance,

postapproval studies, registries etc) and where case-by-case consideration

comes into the decision.


Amgen, Inc AMGEN

7. POSTMARKET SAFETY MONITORING CONSIDERATIONS

7.1 Biosimilar Labeling

Amgen agrees that biosimilar labeling “should include all the information necessary for a

health professional to make prescribing decisions,” as noted in the draft Scientific

Considerations Guidance (lines 821 - 822). In addition to noting whether a biosimilar has

been deemed interchangeable (lines 826 - 827), the labeling should:

• Include data from studies of both the reference product and the biosimilar product

(where applicable), on an indication-by-indication basis, and distinguish the

studies conducted with the biosimilar from those conducted with the reference

product;

• Reflect the outcomes of postmarket commitments and any subsequent clinical

studies, as this information becomes available; and

• Adhere to a standard format.

First, FDA should not simply use the reference product labeling for the biosimilar, as it

does for generic drugs.2

Moreover, these data should be presented on an indication-by-indication basis, so that

prescribers can understand the relevance of available data to the patient population at

issue. This transparency will allow the physician to make an informed choice for each

patient based on his or her particular medical condition, history, and life circumstances.

Instead, biosimilar labeling should include reference product

data and biosimilar data (to help the prescriber understand the database supporting

approval of the biosimilar). The source of the data should be transparent to promote

understanding of the relationship between the products and the available data.

Second, it is essential that biosimilar labeling be kept current. As explained below,

postmarket data will likely be an important source of product safety and efficacy data for

biosimilars. This is because these products are expected to be approved based on

fewer clinical data than their reference products and because premarket clinical trials are

unlikely to study all indications or be able to detect rare adverse events. To ensure that

prescribers have the most up-to-date, comprehensive information about a biosimilar’s

2 FDA’s approach to generic drug labeling is also inapposite for another reason. There is no statutory “same labeling” requirement for biosimilars as there is for generic drugs.


Amgen, Inc AMGEN

safety and efficacy, postmarket data should be promptly added to the labeling as it

becomes available.

Third, Amgen agrees with FDA’s statement in the draft Scientific Considerations

Guidance that biosimilar labeling should note whether FDA has deemed the biosimilar to

be interchangeable with its reference product (lines 821 – 827). This is an essential fact

that could have implications for patient safety, and we support FDA’s plan to clearly

convey this information to both providers and patients. This information will also help

communicate that non-interchangeable biosimilars are not generic products and are not

substitutable for their reference products. We note that it is inconsistent with the concept

of interchangeability, inconsistent with the statutory requirement of “the same expected

result in any given patient” – and impractical for purposes of implementation – to

designate a product interchangeable for less than all indications and routes of

administration.

Finally, FDA should adopt a standard approach to the content and format for biosimilar

labeling. A standard approach will ensure that prescribers can readily locate key

information.

A description of the pivotal study(s) used for the approval of the biosimilar or

interchangeable product should be included in the “Clinical Studies” section of the

prescribing information, along with information on the clinical studies supporting approval

of the reference product. This section should clearly identify the product(s) involved in

each study described. The labeling should cite the safety profile of the reference product

when appropriate, and explicitly acknowledge differences between the profiles of the

biosimilar and reference products when supported by an adequately designed study

(recognizing that the differences should not be clinically meaningful at the time of

biosimilar approval in order to fit within the statutory definition of biosimilarity).

In the adverse event section, the biosimilar labeling should appropriately highlight any

notable adverse events that occur with the biosimilar product but not the reference

product (recognizing that the differences should not be clinically meaningful at the time of

biosimilar approval in order to fit within the statutory definition of biosimilarity), and this

labeling should not suggest a class effect (unless warranted). The “Indications and

Usage” and the “Warnings and Precautions” sections should include only those


Amgen, Inc AMGEN

indications for which the biosimilar has been approved. No reference should be made to

indications that have not been approved or evaluated for approval.

7.2 The Need for Accurate Attribution of Adverse Events and Distinguishable Names

Amgen agrees that there is a pressing need for a robust postmarket safety monitoring

system that distinguishes adverse events associated with the biosimilar from those

associated with the reference product in order to permit detection and assessment of

product-specific safety signals (draft Scientific Considerations Guidance lines 809 - 813).

We encourage FDA to issue guidance on pharmacovigilance issues for biosimilar

products. This guidance should provide recommendations for effective

pharmacovigilance. We strongly recommend that it also state that FDA will require

biosimilars to bear distinguishable nonproprietary names. Because nonproprietary

names may be assigned to proposed biosimilars long before the corresponding

marketing applications are submitted, FDA should make guidance on this topic a high

priority.

7.2.1 Special Considerations for Pharmacovigilance in the Biologics Context

For three reasons, effective pharmacovigilance is particularly important in the biologics

context. First, due to the complexity of these products and the limitations of current

scientific methods, at the time of initial approval manufacturers and regulators often have

incomplete knowledge of a product’s distinguishing quality attributes. They also have

limited knowledge of the relationship between these attributes and a product’s safety

profile. Key safety attributes often must be discerned through postmarket surveillance or

postmarket studies. For example, rare but significant immunologic effects may not be

detected in conventionally sized premarket clinical trials. Moreover, small postmarket

process changes may have significant safety implications, increasing the need for a

pharmacovigilance system that covers the entire product life cycle.

The European experience with Eprex (epotein alfa) and incidences of pure red cell

aplasia (PRCA), discussed in our 2010 comments to FDA, illustrates these features of

biologics and highlights the importance of an effective pharmacovigilance system. In

that case, an unexpected number of cases of a severe immunogenic effect known as

PRCA were reported. Even with only three erythropoiesis-stimulating agents approved


Amgen, Inc AMGEN

in Europe at the time, traceability to the responsible product took months of forensic

research. As a result of this research, the increased rate of PRCA was eventually

attributed to a formulation change for that product (ie, replacing human serum albumin

with sorbitol-80 as a stabilizer) (Kuhlmann et al, 2010). With the arrival of biosimilars,

there will be more products per class on the market, and more challenge in assuring

accurate attribution of events.

Second, biosimilars will inevitably differ from their reference products, and these

differences may have significant safety implications. For example, differences in quality

attributes may lead to differences between products with regard to dose, PK,

immunogenicity, and adverse events (including both serious adverse events, such as

formation of neutralizing antibodies to endogenous mediators (eg, PRCA) and targets,

and less serious adverse events, such as local irritation). Although premarket testing

may help elucidate these differences, some differences will not be detected until after

approval. The Avonex® (interferon β1a) example illustrates this point. In this case,

Bioferon produced the drug used in clinical trials. Bioferon ceased manufacturing for

financial reasons and its partner, Biogen, subsequently began manufacture of the

product using a different cell line.

Biogen sought approval based on the clinical study data generated with the Bioferon

product and a demonstration that the two products were comparable. After two

attempts, Biogen was able to show its product was comparable with the Bioferon

product. Physiochemical testing showed that the products' structures were similar,

multiple bioassays indicated that the products had comparable bioactivity, and PK data

indicated that their distribution and clearance were comparable. FDA therefore approved

Avonex®. Post approval, however, an immunogenicity study showed that the Biogen

product was less immunogenic, indicating that the two products were not as similar as

originally thought (Woodcock et al, 2007). Even though decreased immunogenicity was

advantageous, it represented an unanticipated change not predicted from analytical

tests, bioassay tests, and PK tests.

Third, switching patients between biosimilars and reference products presents special,

as yet not fully understood, challenges. As noted in our 2010 comments, such switching

may cause adverse reactions, particularly immunogenic reactions. In general,

immunogenic responses to biologics are complex and may vary widely with regard to

timing and mode of expression following initiation of treatment. They can range from


Amgen, Inc AMGEN

acute anaphylactoid infusion reactions to delayed development of neutralizing

antibodies. Switching or alternating a patient between a reference product and biosimilar

versions increases concerns regarding the potential for immunogenic side effects. As

Dr. Woodcock has stated, “there is a significant potential for repeated switches between

[proteins] to have a negative impact on [product] safety and/or effectiveness.”

(Woodcock, 2007). Switching also may complicate identification of the causative agent

for a particular adverse reaction for the following reasons:

• Adverse reactions may occur months (or longer) after initiation of treatment with a

biologic, thus making it difficult to attribute a single adverse event to the proper

biologic. Timely reporting of these events, accurate attribution, and signal

detection analyses are critical to effective attribution.

• Infusion reactions can manifest in differing patterns. In some cases, severe

reactions occur initially and then wane with repeated doses. In other cases,

hypersensitivity may manifest after several doses. Because adverse events may

manifest after a latency period, switching may complicate proper attribution.

• An “anamnestic” antibody response to a biologic based on prior exposure may

occur long after the causative agent was administered and only on re-

administration of the product or administration of a biosimilar.

• Due to the nature of an immune response, a stimulus provoked by one biologic

may manifest as a more severe reaction to another, subsequently administered

pharmacologically similar biologic.

For these three reasons, adverse events must be accurately linked to the correct

product, manufacturer, and lot number. A lack of traceability at best leaves unanswered

questions about product safety and at worst misrepresents the safety profile of multiple

products (Miletich, 2010). Unless an adverse event report notes the precise product

involved, adverse events associated with different products will be pooled, potentially

masking problems associated with a particular product and delaying or precluding

appropriate action by regulators and manufacturers.

Designation of a product as interchangeable will create additional challenges for tracking

and tracing and will make accurate attribution of adverse events more important.

Products deemed interchangeable are, by virtue of the designation, more likely to be


Amgen, Inc AMGEN

switched during a course of treatment. In light of the potential for delayed immune

response it will be difficult, if not impossible, to determine the cause of adverse events

that occur after a product switch. Furthermore, rare adverse events may not be

observed in a typical clinical study for a biosimilar product, therefore post-marketing

observation will likely be an essential tool in determining the reporting rate of more rare

adverse events. Product switching could well confound postmarketing assessment of

such semiquantitative data. How to ensure accurate attribution of adverse events

following a product switch remains an open question and should be addressed in future

guidance documents prior to designation of interchangeability.

7.2.2 Suggested FDA Actions

Although no single action by FDA will ensure accurate attribution of adverse events,

Amgen recommends the agency take the following steps.

• Require that biosimilars have distinguishable nonproprietary names,

• Require biosimilars to bear distinctive packaging compared with their reference

products, and

• Take steps to improve the quality of adverse event reporting.

First, biosimilars should have distinguishable nonproprietary names that share a

common root with the reference product’s nonproprietary name but also include a

distinguishing suffix. The use of a common root will allow practitioners and patients to

recognize that the products are related, while the suffix will ensure that the products are

distinguishable.

Distinguishable nonproprietary names will simplify attribution of adverse events to the

correct biologic. Pharmacovigilance depends on voluntary reporting of adverse events

by healthcare providers and patients. Amgen has found that healthcare providers and

patients often do not have or report information needed to permit identification of the

responsible product (eg, lot number, national drug code (NDC)) beyond the

nonproprietary name. For example, NDCs generally are not found on dispensed

pharmacy prescriptions, as FDA’s MedWatch form 3500 instruction notes. Because the

nonproprietary name is often the only product-identifying information on an adverse

event report, the use of a single nonproprietary name for a biosimilar and reference

product will likely confound attribution of the adverse event and overall


Amgen, Inc AMGEN

pharmacovigilance efforts. In contrast, distinguishable nonproprietary names will ensure

that practitioners and patients are provided with the necessary information to accurately

identify the product involved in an adverse event.

Distinguishable names also will provide needed transparency to prescribers and

patients. As discussed in Amgen’s 2010 submission on the Part 15 hearing docket,

neither biosimilar nor interchangeable products are expected to be the “same as” their

reference products, as is the case for generic drugs. Distinguishable names will signal

this fact to the public and will also serve to educate the public that biosimilars, are not

generic drugs. Moreover, the shared root will convey that the products are related and

thus will appropriately reflect that the biosimilar has been shown to have no clinically

meaningful differences from the reference product.

Second, FDA should require biosimilars to have distinctive packaging in all feasible

layers (eg, from external package to syringe or vial labeling). Like distinguishable

nonproprietary names, distinctive packaging will help patients and doctors identify the

product associated with an adverse event.

Third, FDA should take steps to enhance the quality of adverse event reporting. We

suggest that FDA do the following:

• Establish simplified reporting mechanisms (eg, peel-off labels to place in a

patient’s medical record).

• Encourage sponsors to maintain and use centralized systems for collecting and

analyzing adverse event data, so that they can optimize and standardize signal

detection and thus facilitate comparisons between products.

• Educate the public on the importance of accurate adverse event reporting and

the available mechanisms for reporting adverse events.

• Emphasize the importance of obtaining healthcare provider confirmation of any

patient report, particularly for serious adverse events and regardless of whether

the manufacturer or agency receives the event report.

• Encourage sponsors to conduct antibody testing when investigating immunogenic

adverse events and include the results in the adverse event report.


Amgen, Inc AMGEN

8. OTHER ISSUES

8.1 Publicly Available Information

FDA's proposed Q&A Guidance regarding what constitutes "publicly available

information" from the prior approval of the reference product is incomplete. Moreover, its

framing avoids entirely the issue as to what information with regard to a reference

product FDA can rely upon in reviewing and approving a 351(k) application.

The draft Q&A Guidance is incomplete in that it fails to address the issue of what

"publicly available information" FDA will make available as to an approval of a 351(a)

application (ie, reference product) that may be included in a 351(k) application.

Furthermore, while the Agency states that "publicly available information" about a

reference product will comply with the requirements for an "Action Package for

Approval,” as found in Section 505(l)(2) of the Federal Food, Drug, and Cosmetic Act

(FFDCA), it does not confirm that it will issue such approval packages for 351(k)

applications. Since Section 505(l)(2) applies to all 351 applications, FDA should affirm

that 351(k) approval packages will contain the same information as the approval

package for a 351(a) approval and what information, in both cases, that will be. That is,

FDA should require consistency and parity in what is included in approval packages for

both reference products approved under 351(a) and biosimilars approved under Section

351(k).

In all cases, first and foremost, FDA needs to make clear that, regardless of whether the

information pertains to a biological product approved under the (a) or (k) pathway, the

Agency will comply, and has complied, with all legal obligations, including the

requirements and procedures set forth in 21 C.F.R. Part 20, pertaining to the protection

of confidential commercial, trade secret, or otherwise protected information.

The proposed Q&A as set forth in the draft Q&A Guidance is also incomplete in that it

states "the publicly available information posted by FDA in the context does not

necessarily include all of the information that would be otherwise disclosable in response

to a Freedom of Information Act request" (draft Q&A Guidance, p. 11). FDA should

clarify what additional information it suggests might otherwise be disclosable.

Importantly, in all cases, FDA must put appropriate safeguards in place to ensure that it

does not rely upon or otherwise use the confidential trade secret or any other protected


Amgen, Inc AMGEN

proprietary information of the reference product sponsor to review and/or approve a

biosimilar application.

Moreover, we would ask that FDA consider the need to update its BLA disclosure

regulations set forth in 21 CFR. § 601.51 to reflect the existence of a biosimilar pathway.

At a minimum – and more closely aligned with FDA’s longstanding new drug application

disclosure regulations – 21 C.F.R. § 601.51 should be revised to provide that, safety and

effectiveness data and information may not be disclosed at any time should

extraordinary circumstances exist.

Finally, as noted, this Q&A is couched so as to avoid the critical issue as to what

information constitutes "publicly available information" that FDA can rely upon in

approving a 351(k) application. It is a critical issue not only because there are potential

legal issues as to what can be properly used, but due to the fact that a 351(k) application

will be reviewed by the same division that has reviewed the BLA for the reference

product. The Agency needs to set forth clear guidance and rules to the effect that

reviews of 351(k) applications can only rely upon:

• data provided by the 351(k) applicant derived from its own research and

development;

• data to which it has a proper right of reference; and

• data which is properly within public domain (ie, “publicly available").

8.2 Interchangeability

We agree that interchangeability is a distinct and higher standard than biosimilarity. We

appreciate the thoughtful consideration FDA is giving to this important topic. We ask

that, if the agency develops guidance on this topic, it keep the following points in mind.

First, interchangeability of biosimilars is an issue of first impression and warrants a

cautious approach by the agency. The US is the only jurisdiction in which the regulatory

authority must determine whether a biosimilar may be substituted for its reference

product without the involvement of the prescribing physician. It is important that the

pathway for biosimilar products be implemented and regulators evaluate the long-term


Amgen, Inc AMGEN

safety and effectiveness of biosimilars before entering uncharted waters of declaring

biological products to be interchangeable.3

Second, FDA should expand upon its point that interchangeability is a higher standard

than biosimilarity. Specifically, FDA should re-affirm that meeting the standard for

biosimilarity is necessary, but not sufficient, to receive an interchangeability designation.

The statute mandates this approach, by providing that applicants must make two

additional showings, beyond a basic showing of biosimilarity, to receive an

interchangeability rating. FDA should also state that it will designate a product as

interchangeable only if the applicant has obtained approval for all reference product

indications. The statute also requires this approach, by providing that FDA must find that

a product can be expected to have the same clinical result “in any given patient” before

rating a product as interchangeable. Finally, because interchangeability is a higher

standard than biosimilarity, FDA should emphasize in product labeling and educational

campaigns that non-interchangeable biosimilars should not be substituted for their

reference products without physician consent.

Third, a system that provides effective tracking and tracing of all biologic medicines is

necessary to ensure accurate attribution of adverse events caused by biologics. This

system will be increasingly important as more biologics, including biosimilars and

interchangeable products, enter the market and increase the number of products that

could be the source of the adverse reaction. We will only know how safe and effective

products are if we can identify both when an event occurs and the products to which the

patient was exposed. The track and trace system should be in place prior to the arrival

of biosimilar medicines in the US marketplace.

Fourth, the dynamic of the health care market today makes it highly likely that patients

with chronic conditions using biologics will be exposed to multiple products. As the

BPCIA has introduced the legal concept of an interchangeable biosimilar, it is important

to emphasize that non-interchangeable biosimilars are not be substituted without

physician consent. The FDA has an opportunity to clearly specify in the guidances that

the standards established for biosimilarity are separate and distinct to those that are

necessary to secure approval as an interchangeable biosimilar. Therefore, meeting the

3 Although patients might, as a practical matter, be exposed to multiple non-interchangeable products in the marketplace, this experience will not be sufficient to meet the rigorous statutory standard of interchangeability.


Amgen, Inc AMGEN

standard for biosimilarity is necessary, but not sufficient to meet statutory definition of

interchangeability. Due to the importance of this issue we support FDA’s stance that

labeling conventions should clearly state whether or not the drug is “interchangeable”

and a statement on the inappropriateness of interchange when a biosimilar is not rated

as such.

8.3 Sample Retention

In the draft Q&A Guidance, the Agency has proposed that sponsors of 351(k)

applications retain reserve samples of biologic products used in comparative

pharmacokinetics ("PK") or pharmacodynamic ("PD”) studies in accordance with the

requirements of 21 C.F.R. §§320.28 and 320.63. (See A.1.10). Amgen agrees with the

importance of reserve sample requirements for PK and PD studies, but believes it is vital

that rules be in place that specifically apply to 351(k) applications. As conceded by the

Agency, the rules found in 21 C.F.R. §320 only apply to products approved under

Section 505 of the FFDCA. Amgen does not believe it is sufficient for there to be merely

a recommendation to follow the referenced rules; the Agency should, instead, address

the issue through notice and comment rulemaking, with the goal being a rule having the

full force and effect of law, specifically targeted to the unique characteristics of biologic

products.

Reliance upon the rules for sample retention of bioequivalence studies of generics of

drugs approved under Section 505(b)(2) and 505(j) may not be appropriate in all cases

for biologic products. For example, there may in many cases be issues with the integrity

of biologic products stored for as long as five (5) years. A rule is needed to address the

unique stability issues with biologic products. In addition, it may be appropriate to limit

the number of reserve samples that are needed to test for identity and quality. Given

that there are not compendial requirements for most biologics, and that they may not

remain stable, testing for purity may not provide any meaningful data to FDA in terms of

ensuring the integrity of the PK and PD studies, while imposing potentially burdensome

retention programs in terms of both number of samples needed and potentially invalid

testing.

In short, there are differences between simple micro-molecule chemical products

approved under Section 505 and biologic macro-molecule products approved under

Section 351 of the PHSA that render the requirements in 21 C.F.R. § 320 infeasible and


Amgen, Inc AMGEN

inappropriate for biologic products. As there are unique considerations relating to PK

and PD studies of recombinant biologic products and samples to ensure the integrity

thereof, Amgen believes it is necessary for separate binding rules – developed with, and

informed by, stakeholder input through notice and comment rulemaking – be put in place

for reserve sample keeping for PK and PD studies of biologic products.

In the interim, Amgen believes that biosimilar applicants should be responsible for

obtaining FDA acceptance of sample retention plans for PK and PD studies, and that the

Agency should ensure parity across sponsors in terms of the requirements imposed.

8.4 Reference Product Exclusivity

Amgen encourages FDA to continue reviewing the reference product exclusivity

provisions of Section 351(k)(7) of the PHS Act, and appreciates the Agency’s public

acknowledgement that it is doing so (FDA draft Q&A Guidance at 15, acknowledging that

it is continuing to consider the input received to date, including from the 2010 Part 15

hearing docket (Docket No. FDA-2010-N-0477)). Notably, there are several issues that

remain unresolved, which the draft Q&A Guidance does not address. Indeed, there is

only one draft Q&A provided for reference product exclusivity, and even that is cursory

and raises several concerns.

Specifically, the one question posed in the draft Q&A Guidance is whether an applicant

can include a request for reference product exclusivity in its BLA (351(a)) submission.

FDA, in its proposed answer, confirms that the applicant may do so and further advises

that such a request for exclusivity describe how the proposed product meets the

statutory criteria for exclusivity, including that the applicant provide adequate supporting

data and information. In other words, FDA’s proposed response places the burden onto

the applicant to request and, in essence, “make its case” for exclusivity. Such burden-

shifting, however, is contrary to the plain language of the statute.

Sections 351(k)(7)(A) and (B) are drafted so as to confer a presumption of exclusivity for

a stand-alone BLA filed under 351(a) that serves as the Reference Product for a

biosimilar applicant. Specifically, Section 351(k)(7)(A) (“Effective Date of Biosimilar

Application Approval”) dictates that approval of a biosimilar application may not be made

effective for 12 years after the date that the Reference Product was first licensed. In

similarly crafted clear language, Section 351(k)(7)(B) (“Filing Period”) provides that a

biosimilar applicant may not even be submitted until four years after the date that the


Amgen, Inc AMGEN

reference product was first licensed. In other words, there is an affirmative presumption

of exclusivity.

The last provision in Section 351(k)(7) -- Section 351(k)(7)(C) (“First Licensure”) – then

identifies the exception to the rule. That is, Section 351(k)(7)(C) describes those

circumstances under which a separate grant of exclusivity would not apply to a

subsequent application from the same sponsor or manufacturer. There is nothing in this

or any other provision that suggests any intent to have the applicant proactively assert its

right to a 12-year data exclusivity period. In addition, not only does FDA advise the

applicant to justify a claim for exclusivity, it offers no guidance whatsoever on the data

and information that be provided to do so. Given that neither of the statutory provisions

conferring the legal right to exclusivity provides any sort of “inclusion” criteria, it is

unclear what sort of justification the applicant would be compelled to provide in order for

exclusivity to be deemed by FDA to be merited. Ostensibly, then, the sponsor is left with

defending what appears to be a presumption on FDA’s part that unless demonstrated

otherwise, the first-licensure provision codified at Section 351(k)(7)(C) is presumed to

apply; in other words, the applicant is presumed ineligible for exclusivity unless and until

demonstrated otherwise. This turns the reference product exclusivity provision on its

head.

Finally, the proposed exclusivity Q&A is also of concern in that the applicant is instructed

to include its request for exclusivity in its BLA. Although not entirely clear, this raises the

concern that the sponsor would not have any insight into FDA’s thinking as to whether

the product at issue will be awarded 12 years of exclusivity. Indeed, based on this

proposal, it is conceivable that the applicant may not learn of FDA’s decision on

exclusivity until the time of the approval of the BLA itself. In any case, a decision as to

whether a product can be expected to be accorded exclusivity is needed much earlier in

the development process. Exclusivity considerations are a key component of investment

decisions. Uncertainty about whether a product will be eligible for data protection will

significantly deter investments that improve products and will severely impede biotech

development and therefore medical advances for patients.

Thus, we urge that FDA implement a process that allows a BLA sponsor to engage with

FDA at any time in the development process in order to get clarity around whether there

are any issues presented that may affect the grant of exclusivity. At a minimum, for

example, FDA should engage with the sponsor as early as possible to address any


Amgen, Inc AMGEN

“relationship of entity” issue described in the so-called “first licensure” provision (Section

351(k)(7)(C). This sort of issue is an example of one that does not require any sort of

scientific assessment to determine, and thus should be able to be very fully addressed

separate and apart from review of the BLA itself.

In terms of the reference product exclusivity issues on which the draft Q&A Guidance is

completely silent, Amgen would refer the Agency to its written comments submitted to

FDA’s 2010 Part 15 hearing on implementation of the BPCIA rather than recite them

again here. (Amgen written comments dated December 29, 2010, submitted to Docket

No. FDA-2010-N-0477.) In brief, however, Amgen would like to reiterate the following

points:

• Any difference in the structure of a product resulting in a change in safety, purity,

or potency from a previous product is statutorily entitled to its own 12-year

exclusivity period. The exclusivity grant provides an important incentive to

ongoing research and development. It achieves the sought after balance

between encouraging innovation while at the same time facilitating competition.

• The statutory terms “same sponsor manufacturer,” “licensor,” and “predecessor in

interest” are clear; “other related entity” should be read together with these terms

to effectuate congressional intent, and should be interpreted narrowly so as to

not inadvertently stifle innovation that leverages these collaboration efforts.

Moreover, it is clear from the statutory scheme that “related” ties back to the

reference product in question and would not encompass other types of unrelated

“relationships.” FDA should abide by this plain meaning in its implementation of

the BPCIA.

8.5 Definitions

8.5.1 Protein

At this early stage of BPCIA implementation, Amgen concurs with FDA’s proposed focus

on polymer size to define “protein.” That said, Amgen also recognizes the potential need

for flexibility as FDA continues to evaluate whether other considerations may be

appropriate.


Amgen, Inc AMGEN

9. REFERENCES ICH of Technical Requirements for Registration of Pharmaceuticals for Human Use. Guideline S6: Preclinical Safety Evaluation of Biotechnology-derived Pharmaceuticals (R1). June 2011. United States Food and Drug Administration. Guidance for Industry: Population Pharmacokinetics. February 1999. http://www.fda.gov/downloads/ScienceResearch/SpecialTopics/WomensHealthResearch/UCM133184.pdf. Accessed 22 March 2012. United States Food and Drug Administration. Guidance for Industry: Scientific Considerations in Demonstrating biosimilarity to a Reference Product. Draft. February, 2012. Safe and Affordable Biotech Drugs: The Need for a Generic Pathway: Hearing Before the H. Comm. on Oversight and Government Reform, 110th Cong. 33 (2007) (statement of Janet Woodcock, M.D., then Deputy Commissioner and Chief Medical Officer, FDA).

http://www.fda.gov/downloads/ScienceResearch/SpecialTopics/WomensHealthResearch/UCM133184.pdf�

http://www.fda.gov/downloads/ScienceResearch/SpecialTopics/WomensHealthResearch/UCM133184.pdf�