By James DeGiulio --
In an article published in the New England Journal of Medicine on August 4, U.S. Food and Drug Administration (FDA) executives presented a sneak preview of what to expect when the FDA issues its biosimilar standards under the Biologics Price Competition and Innovation Act (BPCIA) later this year. The article, entitled "Developing the Nation's Biosimilars Program," was authored by Dr. Steven Kozlowski (Director, Office of Biotechnology Products), Dr. Janet Woodcock (Director, Center for Drug Evaluation and Research), Dr. Karen Midthun (Director, Center for Biologics Evaluation and Research), and Dr. Rachel Behrman Sherman (Associate Director for Medical Policy).
If the article was intended to indicate how far the FDA has progressed in formulating biosimilar standards, it appears we may be waiting a while longer, as the article provides minimal substantive guidance. However, if a preview of substantive biosimilarity standards is desired, the article does indicate that the FDA is evaluating the adoption of some of the guidelines previously developed by the European Medicines Agency (EMA). The article makes several references to the EMA's Guideline on Similar Biological Medicinal Products, published in 2005, as well as the EMA's 2010 Guideline on Similar Biological Medicinal Products containing Monoclonal Antibodies (see "EMA Publishes Guidelines for Biosimilar Antibodies - Part I" and "Part II"), answering the question of how much the FDA would seek to harmonize its biosimilar guidelines with those of the EMA, and suggesting that the EMA Guideline documents could provide a substantive sneak preview of the FDA's biosimilar standards.
Rather than provide any details on the actual scientific standards, the authors instead categorize the FDA's tentative approach to formulating these standards using a series of buzzwords: (1) "totality of the evidence," including sponsor input on whether further clinical studies are needed; (2) no chance of a "one size fits all" standard akin to bioequivalence/bioavailability under the Hatch-Waxman ANDA scheme; and (3) depending on how well the mechanism of the biologics are characterized, "fingerprint identification" studies, rather than full scale clinical trials, may be permitted to establish biosimilarity with a reference product.
Biologics vary in complexity from cellular therapies to small, highly purified proteins, although even the simplest biologics typically are more structurally complex than the small-molecule drugs in today's market. Partly due to this complexity, new biologics require a substantially greater investment from BLA filers (or "sponsors") for their development and regulatory approval. As a result, biologics tend to be more expensive than small-molecule drugs for patients, and the article notes that these cost obstacles currently limit patient access to these biological products. The Hatch-Waxman framework has proved very successful in overcoming cost barriers and improving patient access to small-molecule drugs. In 2009, almost 75% of small-molecule prescriptions dispensed in the United States were for generics, and the approval of a generic drug resulted in average savings of 77% of the product's cost within 1 year. The article acknowledges that cost reductions for biosimilars probably won't be as large, at least in the early going, but points to the Federal Trade Commission report ("Emerging Health Care Issues: Follow-on Biologic Drug Competition"; see "No One Seems Happy with Follow-on Biologics According to the FTC"), which predicts that the availability of biosimilar products will significantly reduce the cost of biologics and increase their accessibility.
The article outlines the familiar challenges facing the FDA in formulating the standards that an applicant must meet to secure approval for a follow-on biologic product. The agency must establish scientific criteria that address the key question: how similar is similar enough when it comes to the substitution of complex biologic drug products in clinical practice? The complex structures of biologic products are not as easily characterized as small-molecule drugs, with biologics generally requiring more advanced and expensive technology. Further, therapeutic proteins must have a specific set of structural features, such as glycosylation and protein folding, essential to their intended effect, and slight modifications in these features can affect the performance of these drugs in humans. This additional secondary structural layer is required to fully characterize biologics, and this layer is more difficult to detect using conventional technological means. This secondary structural layer, often represented by chemical modifications of the amino acid sequence of the protein, has proven to be just as important as the primary amino acid sequence and structure, for changes in the chemical modifications of biologics affect their immunogenicity and safety profiles.
The authors quickly turn to reassuring the reader that the FDA is fully equipped to deal with these challenges, discussing the agency's expertise in this area. The article notes that the FDA has worked with biologics since the mid-1990s, evaluating the physiochemical and functional assays used to characterize biologics and test changes in manufacturing processes. The FDA has made decisions on when animal or clinical studies are required to resolve any remaining uncertainties about the comparability of the products created before and after such changes, these decisions being extremely relevant in setting the biosimilar standards. The FDA must consider the following balance: the biosimilar applicant must be required to establish with sufficient confidence that safety and efficacy are not diminished with their product. However, requiring the biosimilar applicant to repeat all of the clinical trials conducted by the BLA holder would defeat the purpose of the regulatory framework. The FDA acknowledges that simply repeating all of the sponsor's trials is not feasible and is too costly. Utilizing knowledge from the reference products, the abbreviated pathway must eliminate unnecessary (and therefore unethical) testing of biosimilars in animals and humans.
Despite the FDA's experience in deciding when additional clinical trials are required to establish similarity between two products, the authors prudently acknowledge that the EMA has even more experience in this area. The EMA biosimilar guidelines have been in place since 2005, and the EMA first approved a biosimilar in 2006. The FDA seems to be fashioning its approach after the EMA, relying on the EMA's more than five years of experience in implementing a biosimilar framework, and latching onto the EMA's standards and ideas for determining exactly how many studies will be required.
Continuing with the theme of acknowledging others with more experience than the FDA, the article suggests that BLA holders should expect to be heavily involved with setting biosimilar standards and approving biosimilar products than under the Hatch-Waxman framework. Sponsor input will be more extensive and will carry more weight, and extensive product review will be required early on. The new pathway will require a new paradigm for sponsor–FDA interactions. Although the agency frequently meets with sponsors before they submit investigational new drug (IND) applications, a more extensive product review will be required to determine how much additional data and studies are needed for a biosimilar. The first step will be taken by the FDA, which will conduct an in-depth review of the biosimilar applicant's comparative analytic characterization and in vitro data. Once the review of this data is completed, the FDA will work with the BLA holder/sponsor to determine the scope of any required animal and human studies. Since the BLA holder/sponsor has an interest in keeping the biosimilar product off of the market, the agency is currently considering how to structure this interaction so that neutral and more reliable input can be acquired.
The FDA has traditionally relied on integrating various kinds of evidence in making regulatory decisions, and drafting the biosimilar guidelines will be no exception, as seen by importing the EMA and BLA holder standards. The FDA authors prefer a flexible "totality of the evidence" approach for assessing biosimilars, since multiple assay technologies are currently used to characterize the structural attributes of biosimilars, and more likely to be developed going forward. For small molecules, a generic's structure is readily determined by nuclear magnetic resonance, mass spectrometry, infrared spectrometry, X-ray crystallography, or other well-known physical methods. However, accepted methods such as these do not yet exist for biosimilar products. Developers of biosimilars who intend to match reference products more closely will need to select appropriate assays and source materials, and tune their processes carefully. Current approaches to manufacturing process design and improvements in process analytics will become more useful as the technology continues to progress.
Even if certain technological assays become recognized as the gold-standard for determining a biological structure, given the complex nature of biologics, it's unlikely that a "one size fits all" systematic assessment of biosimilarity will ever be developed. Indeed, the EMA biosimilar guidelines are highly product-specific, with the EMA guidance providing structural, animal, and clinical study requirements for each biosimilar product. This is in sharp contrast to the Hatch-Waxman small molecule scheme, where often an ANDA filer must show only bioavailability and bioequivalence for approval. In the biosimilar setting, FDA scientists will need to integrate multiple types of information from multiple assays to provide an overall assessment that a biologic is biosimilar to an approved reference product. The authors also suggest that implementing a more rigorous pharmacovigilance program will also be significant information source, with an emphasis on tracking adverse events and even the smallest change in manufacturing. This will inform future decisions not only on that particular product, but also on products that can be grouped in the same class, such as monoclonal antibodies.
Importantly, the development of the mechanism of the biologic's activity will also factor in determining the product-specific standards for the particular biosimilar. If the mechanism is well known, there might be more flexibility to deviate from the structure of the approved reference product. For example, if the literature shows that all known glycosylation patterns of a biologic molecule show comparable activity and immunogenicity, a glycosylation pattern identical to the reference product may not be required for a finding of biosimilarity. Understanding the mechanism of biologic activity, including any clinical information available, will help evaluate the risks of changes in manufacturing by the applicant and the resulting potential to introduce harmful impurities. This factor is significant to the basic science research community, including university research labs, which could be in a position to secure funding from biosimilar applicant companies that need to establish the safety profile of new manufacturing protocols.
If the mechanism of a biosimilar product is well-characterized, the FDA may accept a "fingerprint"-like identification of similar patterns in two different products. The EMA monoclonal antibody guidelines introduce an outline to consider when designing biosimilarity studies, including the use of populations, pharmacodynamic markers, and end points that are sensitive to the potential differences between products. The EMA guidelines allow a biosimilar monoclonal antibody to differ structurally from the reference product, providing a specific set of "fingerprint" metrics are satisfied. These "fingerprint" strategies were used to secure the approval of the generic low-molecular-weight heparin product enoxaparin, which is structurally different from the reference product. The authors admit that although animal and clinical studies will generally be needed for protein biosimilars for the foreseeable future, the scope and extent of such studies may be reduced further if more extensive "fingerprint"-like characterization is used. Therefore, it is conceivable that at some future date, there may be a set of minimum "fingerprint" biosimilar characteristics, akin to the familiar bioequivalence and bioavailability minimum requirements seen in Hatch-Waxman ANDA filings. Of course, that future date is likely decades away.
Finally, the authors seemingly feel obligated to address the "interchangeability" standard of the BPCIA, which provides for when a biosimilar may be substituted for the reference product without requiring the prescriber's intervention (as in the classic generic drug scenario). The language in the BPCIA focuses on alternating or switching between the two products, requiring the biosimilar to provide the same risk and clinical result as using the reference product throughout. The article presents little reason for optimism in meeting such a heightened standard, instead focusing on the FDA's dedication to policing the marketplace to ensure that merely biosimilar products are not substituted for a reference product without the prescriber's consent. Even equipped with the EMA guidelines as a template, the FDA has its hands full in determining what should be the standards for biosimilarity. Until the FDA's guidelines are established, the standard for interchangeability is likely to be left to statutory interpretation according to the language in the BPCIA. This (perhaps purposefully) vague standard is extremely difficult to meet as it reads textually, and it would be surprising if the FDA approves any biosimilar as interchangeable before the year 2020.
Comments