The Industry Weighs in on FDA’s Draft Biosimilar Guidance
BIO, PhRMA, GPhA, EGA as well as Pfizer, Sanofi, Novartis, Genentech, AbbVie, Boehringer Ingelheim, and Novo Nordisk offer input on the FDA’s draft guidance on the clinical studies and related analytical testing needed to support biosimilarity.
The US Food and Drug Administration (FDA) took an important step toward developing a regulatory pathway for biosimilars in the United States with the issuance in May 2014 of a draft guidance, Guidance for Industry: Clinical Pharmacology Data to Support a Demonstration of Biosimilarity to a Reference Product, which describes the clinical pharmacology studies and related analytical testing needed to support a decision that a proposed therapeutic biological product is biosimilar to its reference product. In its final form, the guidance will be one in a series that the FDA is developing to implement the Biologics Price Competition and Innovation Act of 2009 (BPCIA), which was enacted as part of the Patient Protection and Affordable Care Act, to establish an abbreviated pathway for FDA licensure of biological products that are demonstrated to be biosimilar to or interchangeable with an FDA-licensed reference product. The public comment period for the draft guidance ended August 12, 2014. Several industry organizations, such as the Biotechnology Industry Organization (BIO), the Pharmaceutical Research and Manufacturers of America (PhRMA), the Generic Pharmaceutical Association (GPhA), and the European Generic medicines Association (EGA) as well as pharmaceutical companies, including Pfizer, Sanofi, Novartis, Roche/Genentech, AbbVie, Boehringer Ingelheim, and Novo Nordisk, were among those offering comments. So what did the industry have to say? DCAT Value Chain Insights examines the industry’s input.
Understanding the draft guidance
The draft guidance pertains to therapeutic biological products for which pharmacokinetic (PK) and pharmacodynamic (PD) data are required as part of a stepwise approach to developing the data and information necessary to support a demonstration of biosimilarity. Specifically, the draft guidance discusses some of the overarching concepts related to clinical pharmacology testing for biosimilar products, approaches for developing the appropriate clinical pharmacology database, and the utility of modeling and simulation for designing clinical trials. The draft guidance describes three key concepts for the development of proposed biosimilar products: exposure and response assessment, evaluation of residual uncertainty, and assumptions about analytical quality and similarity. The draft guidance also examines the bioanalytical methodology and clinical pharmacology studies needed to gain safety and immunogenicity information.
In the draft guidance, FDA says it will use a risk-based approach to evaluate biosimilarity by considering the totality of the data and information submitted, including, for example, data from structural and functional characterization, nonclinical evaluations, human PK and PD studies, clinical immunogenicity testing, and studies for clinical safety and, when necessary, clinical effectiveness. The draft guidance suggests that the data be collected in a stepwise manner. The need for additional studies at each step will be determined by the degree of “residual uncertainty” that remains at each step regarding the similarity of the products and whether or not the study can address these uncertainties.
In this stepwise assessment of biosimilarity, the draft guidance says that comparative structural and functional studies (e.g. bioassays, binding assays, and studies of enzyme kinetics) should be performed to evaluate whether the proposed biosimilar product and the reference product are highly similar. A meaningful assessment depends on, among other things, the capabilities of available analytical assays to assess, for example, the molecular weight of the protein, its higher order structure and post-translational modifications, heterogeneity, functional properties, impurity profiles, and degradation profiles denoting stability. The draft guidance outlined four possible assessments of biosimilarity (i.e., “not similar,” “similar,” “highly similar,” and “highly similar with fingerprint-like similarity”) (see Table I). The outcome of the comparative analytical characterization would inform the next steps in the demonstration of biosimilarity.
|Table I: Four Assessments of Biosimilarity as Outlined in the US Food and Drug Administration Draft Guidance*
|Not similar||Certain differences in the results of the analytical characterization may lead to an assessment of “not similar” and further development through the 351(k) regulatory pathway (i.e., regulatory pathway for biosimilars) is not recommended unless, for example, modifications are made to the manufacturing process for the proposed biosimilar product that is likely to lead to a highly similar biological product.|
|Similar||Further information is needed to determine if the product is highly similar to the reference product. Additional analytical data or other studies are necessary to determine if observed differences are within an acceptable range to consider the proposed biosimilar product to be highly similar to the reference product. As an example, glycosylation plays an important role in the PK of certain protein products. Manufacturing process conditions may impact glycosylation. Comparative PK and PD studies of the proposed biosimilar product and the reference product help resolve that some differences in glycosylation identified in the analytical studies would be within an acceptable range to consider the proposed biosimilar product to be highly similar to the reference product.|
|Highly similar||The proposed biosimilar product meets the statutory standard for analytical similarity.The results of the comparative analytical characterization permit high confidence in the analytical similarity of the proposed biosimilar and the reference product, and it would be appropriate for the sponsor to conduct targeted and selective animal and/or clinical studies to resolve residual uncertainty and support a demonstration of biosimilarity.|
|Highly similar with fingerprint-like similarity:||The proposed biosimilar product meets the statutory standard for analytical similarity based on integrated, multi-parameter approaches that are extremely sensitive in identifying analytical differences. The results of these fingerprint-like analyses permit a very high level of confidence in the analytical similarity of the proposed biosimilar and the reference product, and it would be appropriate for the sponsor to use a more targeted and selective approach to conducting animal and/or clinical studies to resolve residual uncertainty and support a demonstration of biosimilarity.|
| *US Food and Drug Administration (FDA), Guidance for Industry: Clinical Pharmacology Data to Support a Demonstration of Biosimilarity to a Reference Product (Draft Guidance, May 2014).
Industry organizations offer input
Organizations from both the innovator and generic-drug sides of the industry offered their feedback on the draft guidance. On the innovator side, PhRMA and BIO submitted comments as did GPhA and the EGA on the generic-drug side.
PhRMA offered five key comments. First, it suggested that the FDA remove the four-tiered framework (see Table I) from the draft guidance because it believed it was beyond the scope of the draft guidance and that the definitions of the four categories were “vague and confusing.” If the framework were to be included in the draft guidance or other guidance, PhRMA recommended that the FDA clarify the meaning of these categories and explain how the agency intends to use them. Specifically, it recommended that the FDA explain whether it will quantitatively separate biosimilar candidates into each of the four listed categories and, if so, how it would do so. PhRMA further suggested that all stakeholders, not just those focused on clinical pharmacology issues, be included in offering feedback and that a revised framework be included in other FDA draft guidance, such as Scientific Considerations in Demonstrating Biosimilarity to a Reference Product (February 2012). PhRMA suggested that the agency clarify when it believes that clinical pharmacology studies constitute the full clinical evaluation and, if possible, provide specific examples.
PhRMA also suggested that the FDA should clarify when it believes that clinical pharmacology data is sufficient to demonstrate that a proposed biosimilar has no clinically meaningful differences in safety, purity, and potency from the reference product. It also suggested that biosimilar candidates generally be subject to at least one comparative clinical safety and effectiveness trial with the reference product to establish that there are no clinically meaningful differences in safety, purity, or potency between the biosimilar and reference product. PhRMA also suggested that a premarket evaluation of safety and immunogenicity be required as part of the process in establishing biosimilarity and that biosimilar applicants perform postmarket pharmacovigilance to aid in the detection of rare events.
PhRMA also raised concerns with the draft guidance’s statement that if when clinical pharmacology studies fall outside the scope of predefined limits for PK and/or PD similarity of the reference product, the proposed biosimilar product should not be a candidate for approval under the biosimilar regulatory pathway as defined by Section 351 (k) under the Public Health Services Act as established by the BPCIA. PhRMA requested that the guidance state that a biosimilar applicant cannot rely on clinical testing to justify major analytical differences between the proposed biosimilar product and reference product or that a proposed biosimilar candidate cannot overcome major differences in PK/PD or immunogenicity through clinical effectiveness studies. Lastly, PhRMA proposed that the FDA consider issuing guidance that addresses analytical, preclinical, and clinical issues for individual product classes, something which European regulators have done. “Issuing class guidances would help FDA address testing issues specific to types of proposed products and ensure that FDA is transparent about how these data requirements will vary with product attributes,” said PhRMA in its comments.
As did PhRMA, BIO also expressed concerns with the four-tier assessment of biosimilarity as proposed in the draft guidance (see Table I). It said that the draft guidance is not clear what type or scope of information is needed to differentiate a “similar” molecule from one that is considered “highly similar” or “highly similar with fingerprint-like similarity,” and that it is unclear on what the impact (if any) that these different classifications would have on biosimilar product development.
BIO said that the draft guidance introduces examples where a failure to demonstrate that a proposed biosimilar product is “highly similar” to its reference product at a given stage may be justified with additional data. “BIO requests that the agency convey limiting principles for such scenarios so that there is not the possibility that sequential analytical failures could ultimately be justified by virtue of the results of a clinical comparative safety/efficacy trial, which would expose human subjects to an experimental therapy that had not met the statutory analytical threshold of “highly similar.” It also requested further elaboration of the implications of the “highly similar with fingerprint-like similarity” classification.
Also, like PhRMA, BIO also recommended that class-specific guidance be developed because additional considerations for clinical pharmacology data may be needed to support demonstration of biosimilarity depending on the product class. It, too, pointed to the European Medicines Agency (EMA) issuance of class-specific guidance for biosimilars. “Through this approach, the EMA has encouraged open scientific debate about key biosimilarity issues and has released product class-specific guidance documents that aim to provide transparency and clarity to both biosimilar and innovator sponsors. BIO encourages FDA to consider this approach and to develop vertical guidance for specific classes of products.”
BIO also agreed with PhRMA that safety and immunogenicity testing be required before the approval of a biosimilar. It also added the post-approval pharmacovigilance testing be required. On a different matter, BIO further addressed the concept raised in the draft guidance of using multiple PD markers to address the absence or either a single meaningful robust, and relevant PD biomarker or a PD endpoint that is closely associated with clinical outcomes. “FDA seems to suggest pooling or grouping multiple PD markers or endpoints can overcome these significant limitations. While this approach may have utility, it runs the risk of merely increasing the quantity of data without necessarily improving the quality and interpretability of the results,” said BIO in its comments.
GPhA was supportive of the draft guidance “We would note that the GPhA is confident, given FDA’s many years of experience leading the world on the licensure of originator biopharmaceuticals, that the application of consistent regulatory standards will make these biosimilar[s] as safe, pure and potent as originator biologics. This still allows flexibility as to biosimilar data sets and any given biosimilar sponsor’s scientific justifications for their approaches, just as the case today for innovator biologics.”
While supporting the draft guidance, GPhA also offered specific comments and questions. GPhA suggested the comparative analytical characterization for the four-tier framework for biosimilarity assessment would be more aptly included in the FDA’s draft quality guideline and that it needed further clarification. It also recommend that the FDA further clarify the differences in analytical characterization between a biosimilar candidate and reference product that would disqualify a proposed biosimilar product from being evaluated through the biosimilar regulatory pathway. GPhA further recommended that the FDA re-valuate the portion of the guidance dealing with bioanalytical methods in PK and PD studies.
The European Biosimilars Group (EBG), a sector group of the EGA, characterized the draft guidance as a “very good guidance” and offered specific suggestions for improved clarity. Like others, it recommended that the four-tier assessment (see Table I) be removed from the draft guidance. “While it is appropriate to describe how analytical quality and similarity inform clinical pharmacology studies, the details provided in this section are misplaced in a guidance document on clinical pharmacology,” said the EBG. It recommended that these categories not be used in this draft guidance or other guidance. “The similarity categories appear artificial, are not well-defined, and do not add practical value,” said the EGB and suggested that the FDA adhere to the statutory standard of “highly similar” and to delete all other categories (i.e., “not similar,”similar,” and “highly similar with fingerprint-like similarity”).
In offering general comments, the EBG recommended that the new guidance incorporate the requirements on retaining reserve samples of the biological products used in comparative clinical PK and/or PD studies for a 351 (k) application as initially stated in the draft guidance issued in February 2012, Guidance for Industry on Biosimilars: Q&As Regarding Implementation of the BPCI Act of 2009. Specifically, the EGB suggested that the FDA specify the quantities needed for thorough characterization of the reference clinical batches used in PK/PD studies as part of structural and functional characterization. It added that this assessment should be done by taking into account that often times multiple batches of references are used in comparative PK/PD studies.
Pharmaceutical companies’ feedback
Several pharmaceutical companies offered comments to the FDA’s draft guidance and raised many of the points offered by some of the industry associations. In its comments, Roche/Genentech raised a similar concern as PhRMA that the draft guidance went beyond clinical pharmacology and therefore that the guideline’s title was inconsistent with its content. Like BIO and PhRMA, the company also noted that additional clarity of the four-tier assessment (see Table I) would be helpful and also urged for safety and immunogenicity testing prior to biosimilar approval. The company also noted that the new guidance endorses the use of a “fingerprint-like algorithm” to generate comparative characterization data and that such an algorithm would likely vary between classes of molecules. It said that more detail on what this phrase means or references to case studies where this approach had been used would be helpful.
Sanofi also agreed with others’ positions that immunogenicity testing be performed prior to biosimilar approval. The company also suggested that the guidance provide more specific information, such as examples, to assist in determining clinically meaningful differences between a biosimilar candidate and reference product and how to evaluate these differences.
AbbVie offered three main comments. It agreed with others that safety and immunogenicity testing be performed prior to a biosimilar’s approval and that product-specific guidance be offered. It also suggested that the draft guidance conform to statutory language that a biosimilar product be “highly similar” to the reference product notwithstanding minor differences in clinically inactive components, and if it is not, that the application should not proceed under the biosimilars pathway. It specifically recommended that the FDA delete from the draft guidance that a biosimilar applicant may overcome demonstrated structural differences from the reference product through clinical testing or that a biosimilar product may be approved with analytical data only showing that the reference product and biosimilar candidate are “similar” rather than “highly similar.”
Like others, Pfizer also suggested that the four potential assessments with a development-phase continuum (see Table I) needed further clarification and suggested moving this content to FDA’s draft guidance, Scientific Considerations in Demonstrating Biosimilarity to a Reference Product. It also raised concerns that the step-wise development approach might imply that the legal standard for demonstrating that a proposed biosimilar candidate is “highly similar” to a reference product is flexible. “Surpassing that standard (i.e., highly similar with fingerprint-like certainty) is not a prerequisite to approval but instead a means of providing scientific support for reducing residual uncertainty during development and thereby reducing the level of clinical and other investigation required,” said the company in its statements. “In the same regard, an assessment of the comparative analytical characterization that indicates the proposed biosimilar product is merely similar to the reference product would not meet the statutory standard for registration.” The company further suggested that the agency provide more clarification to elucidate the statutory terminology of what is meant by “highly similar.” Morevoer, the company said that offering clear guidance on what constitutes finger-print like similarity (i.e., how it is defined and achieved) would be helpful.
Novartis also recommended that the analytical section of the draft guidance, which includes the four-tier assessment framework (see Table I) be removed from the draft guidance and placed in the FDA draft guidance, Scientific Considerations in Demonstrating Biosimilarity to a Reference Product. If this section is kept in the guidance, like others, the company suggests that further clarification, including more specific definitions and examples, be included for each category. In a separate comment relating to clinical pharmacology data, Novartis suggested that the FDA should comment on whether publicly available PK data (and/or PD) data bridging US and foreign sourced reference product could preclude the need for a biosimilar sponsor to conduct a bridging study.
Novo Nordisk suggested that the draft guidance provide more detail around the criteria necessary for a demonstration of biosimilarity and reduction of residual uncertainty. To this end, it raised certain questions: “What are the limiting principles for how many failures one can have? While sponsors may have different assays to assess biosimilarity, the guideline could provide the expectations for testing validity of the assay and the requirements for describing limitations of the assay,” said the company in its comments.
On the whole, the company said that the draft guidance could be read to permit unlimited attempts to “prove” a product is a biosimilar.
The company said that the guidance should clarify that a lack of similarity in the quality section of the draft guidance cannot be bridged by clinical studies. It also criticized the draft guidance for weakening the scientific rigor presented in other draft guidance as it relates to criteria for clinical pharmacology and that the draft guidance should establish standards for PK exposure and PD response to set forth expectations for clinical pharmacology. It also said that further clarification is needed since the draft guidance implies that clinical assessment can overcome structural dissimilarities between the reference product and biosimilar produt, which is not allowed under BPCIA.
In its comments, Boehringer Ingelheim points that the draft guidance shifts the responsibility for collection of sufficient data to the biosimilar developer and recommended common ground between originator and biosimilar development. “Even without changing anything, batch-to-batch differences do exist without understanding their influence on clinical efficacy and safety. The originator should define the extent of changes in the substance composition that have a relevant effect on clinical PK, clinical efficacy, receptor binding and intrinsic activity. Although this sounds complex, it is not, and binding studies for each glycosyation form as well as PK studies in Phase I for each glycosylation pattern should be part of biosimilar/originator development.”